Home
Videos uploaded by user “HasGeek TV”
Anirudh Sanjeev - Understand, find, prevent and fix memory leaks in JavaScript
 
19:00
* Do we even need to worry about JS memory. It’s garbage collected after all. * How do JS engines implement Garbage Collection and other memory stuff. * How memory is managed outside the JS world - in Java, C, etc. * Common ways you can leak memory or take up too much of it. * How to use chrome devtools to find memory leaks. * Understanding retaining trees and retaining paths - a core concept a good JS dev should know. https://jsfoo.talkfunnel.com/2015/1-understand-find-prevent-and-fix-memory-leaks-in-ja
Views: 7711 HasGeek TV
Arindam Paul - JavaScript VM internals, EventLoop, Async and ScopeChains
 
38:24
The goal of this talk would be to provide a good understanding of the following, A crisp understanding of the JavaScript VM and how a single threaded engine can be massively parallel. How event loop and callbacks works, example of blocking and non-blocking codes, ES6 generators for custom Async signaling. How function definitions happen and invocations happen, how scope chains are formed, what closures are and how to reason about them. Async programming, how reactive programming works at the bottom level, how event listeners registers themselves and then deregisters. https://jsfoo.talkfunnel.com/2015/44-javascript-vm-internals-eventloop-async-and-scopec
Views: 70816 HasGeek TV
[Flash Talk] Where the fuck is the JavaScript developer?
 
05:09
Harish rants about the lack of pure JS developers with everyone specialising in frameworks.
Views: 10726 HasGeek TV
Siddhartha Reddy - CAP Theorem: You don’t need CP, you don’t want AP, and you can’t have CA
 
43:38
If you use any distributed database, you would have surely heard of the CAP Theorem: “Consistency, Availability, Partition tolerance — choose any two!” But CAP theorem is a little more nuanced than that, and it’s impact on how you should choose and operate distributed databases even more so. Consider these two examples: A distributed key-value store where nodes accept all writes but then respond to all read requests with a null. A distributed key-value store where every node rejects all write as well as read requests. We can build the above systems using just some bash utilities. The first is an example of an AP system and the second is an example of a CP system; in fact, these examples are more-or-less straight out of the paper that proved the CAP Theorem. Given that such trivial systems can be classified as CP/AP, is CAP-Theorem even a viable framework for analysing distributed databases? Tentatively, the talk will cover: intro to CAP theorem (the assertion; meaning of C, A, & P; proof) examples of trivial CP & AP systems why CA is not meaningful, CP is not viable, AP is not useful alternatives to CAP theorem analysis of a few real-world databases
Views: 10899 HasGeek TV
Better Async with Redux Saga - Kushan Joshi, Mapbox
 
21:25
This talk would assume basic familiarity with Redux. I would use half of the talk to explain what are generator functions. Generators are very different from anything javascript has had in the past. This turns off many people to adopt Redux Sagas as each heavily relies on generators. The remaining half I will cover how we handle async actions in projects at Mapbox. The problem with redux-thunk and how Redux saga helps convert async control flow to a more synchronous like control flow. Talk details: https://reactfoo.talkfunnel.com/2017/16-better-async-with-redux-saga ReactFoo 2017 website: https://reactfoo.in
Views: 6348 HasGeek TV
Anand Chandrasekaran - Keeping Moore's law alive: Neuromorphic computing
 
43:02
Moore’s law is a term coined by Carver Mead, a Caltech professor who is also the father of Neuromorphic Engineering. It refers to the observation, now more hope than reality, that advances in technology will allow a doubling of compute capability in silicon every 18 months. Recent advances in the use of highly parallel compute methods, that are loosely based on neural systems in our brain, are changing how compute is accomplished. These techniques, collectively termed deep learning networks, burst onto to the world because of one reason: the ability to perform lots of parallel computations on graphics cards. However, it is in truly custom hardware, such as that pioneered by the Neuromorphic community that we will find the salvation of Moore’s law. When we blend powerful compute techniques with custom silicon architectures, we can keep the hope alive of continuing to double the compute capability of the world. If you are in the space of deep learning or have heard about how GPUs have revolutionalized high performance computing, this talk will take you to the extreme bleeding edge of that world.
Views: 8780 HasGeek TV
Rohit Gupta – Logging at scale using Graylog - Billion+ messages, 100K req/sec
 
21:13
Slides: https://www.rohit.io/res/talks/logging-at-scale-using-graylog-billion-messages-100k-req-sec/ With the advent of micro-services, dozens of releases per day, logs are the bread and butter for a successful real-time technology platform like OlaCabs. In this talk, I would be presenting our logging pipeline and the challenges we faced while doing it at Ola scale.
Views: 3397 HasGeek TV
Scaling NodeJS - Abhinav Rastogi, Flipkart
 
37:36
My journey with Node started over 3 years ago, and I presented my initial learnings about running Node in production at JSFoo 2014. We have been using Node as our primary web serving layer since then, adding more features, more properties and most importantly, more traffic onto it. This led to the discovery of a lot of bottlenecks and a lot of late-night load tests. In this talk, I’ll be focusing on scaling a Node based web server and what kind of issues we have run into. This will include different approaches to attacking this problem like horizontal and vertical scaling. I will be taking you through the variety of resource bottlenecks you can expect to run into, like network, memory, disk and cpu. I’ll touch upon how to find these bottlenecks and what technologies you can use to solve them. Running a Node server in production is not for the faint of heart, but it gets much easier when you know how to juice out the maximum performance from your systems! Talk details: https://jsfoo.talkfunnel.com/2017/98-scaling-nodejs JSFoo 2017 website: https://jsfoo.in/2017/
Views: 6359 HasGeek TV
Functional MVVM using RxJava and Android Data Binding - Manas Chaudhari, TinyOwl/Runnr
 
35:35
- Problems in building User Interfaces - Mutation causes refresh chain - Boilerplate for view setup - Solving Mutation using RxJava - Eliminate View Boilerplate - Problems in binding rx.Observable to Views - Solution using Data Binding - Architecture - Intro to MVVM - View Composition using MVVM - Static composition using include tag - Setting up RecyclerView/ViewPager in XML - Managing Dependencies - Testability -Conclusions https://droidconin.talkfunnel.com/2016/32-functional-mvvm-using-rxjava-and-android-data-bind
Views: 13085 HasGeek TV
Bernd Erk - Why favour Icinga over Nagios
 
45:39
Most sys admins have a love-hate relationship with Nagios based monitoring solutions. Backed by a sizable community, users have learned to live with it’s shortcomings in scaling, configuration, and modern integration options. Taking advantage of the tremendous number of supported hard- and software, Icinga leaves all legacy limitations behind. It delivers an easily scalable solution, with clustering, load balancing, automated replication, and even business process monitoring out-of-the-box. Based on a new configuration format with advanced language features - like conditional processing and complex type support - monitoring agile environments works like a breeze. Existing modules for Puppet, Chef and Ansible ramp up the rollout time and ensure a continuous and up to date monitoring environment. The talk will demonstrate how popular tools such as Graphite, Logstash, or Graylog integrate better and easier than ever before. In addition to that we’ll introduce the new Icinga Web 2 interface and give a brief introduction into the technical architecture. Bernd Erk, is co-founder and project organiser of the Icinga project that has been busy improving the forked Nagios code base since 2009. In his day job as Managing Director of NETWAYS he overseas success and smooth operation of all customer projects and business processes. His technical expertise stretches across Systems Management, Managed Services and Software Development. A contributor to Linux Magazine and Linux Technical Review in Germany, Bernd regularly publishes articles and presents on open source topics ranging across monitoring, virtualization, databases and performance tuning among others. Bernd was previously Operating Systems Specialist at Quelle Schickedanz AG & Co., where he worked heavily with Solaris, HPUX and Oracle databases. After which, Bernd spent 8 years as Business Unit Manager at Ise-Informatik where he dealt with Oracle databases and service oriented architectures.
Views: 15454 HasGeek TV
Saptarshi Prakash, Zeta - UX Choreography: motion design and micro-interactions
 
31:10
This is 2016 and every digital product aims to speak to its users. Be it an app which books your cab or a website that manages your finances - everything attempts to be livelier, engaging and more delightful than the inanimate (text + image) combo. A bit of internet research will reveal that Motion Design had been an unexplored teritory in the past. But now, the tables have turned and people all over the world are busy building lively and ‘cool looking’ products. But how cool should a product be? At times while browsing some visually heavy apps and websites, I get so engrossed in appreciating the mesmerising motions that I ended up ignoring the content. The key is to create the right balance between content and visuals and this is what I m going to speak about, along with an introduction to animations in After Effects and its implementation in CSS. https://metarefresh.talkfunnel.com/2016/21-ux-choreography-motion-in-ui
Views: 3861 HasGeek TV
Weaponizing data for politics - Shivam Shankar Singh
 
30:23
We’ve all heard of the prevalence of Data Analytics in the political realm and stories of how companies like Cambridge Analytica influenced elections with the use of data. It used to be information that was power in politics but now data and the analysis of it lets parties weild even more power. If done right and combined with on ground intelligence, it allows for microtargetting and targetted advertising that are now a major tool for shaping the public discourse. My talk would focus on what data can be used by political parties and consultants to influence elections and how this data is converted into strategy. The talk would also raise some ethical concerns on the use of data for political purposes so that people working in the technology sector can have a more informed discussion on data privacy laws in the country. It would also be a call to action for people in the tech space to raise their concerns about how most of the data is stored in extremely insecure databases in the country. Shivam is a politicial consultant working primarily on the data side and have worked on the Manipur and Tripura Legislative Assembly election campaigns for BJP. Shivam also worked briefly on the Punjab Legislative Assembly elections with Prashant Kishor’s company IPAC and was a Legislative Assistant to Member of Parliament (LAMP) fellow in 2015-16. Shivam graduated from the University of Michigan - Ann Arbor with a B.Sc. in Economics.
Views: 1730 HasGeek TV
Siddhartha Reddy - Build Your Own Search Engine
 
23:55
No, this is not another tutorial on using Solr/ElasticSearch/Sphinx/Lucene. Imagine that none of these existed and you need a search engine for your shiny new eCommerce startup. What would you do? Build your own search engine, of course. In this session at The Fifth Elephant 2012, Siddhartha Reddy describes what it takes to do that. The Fifth Elephant is a conference on the big data ecosystem by HasGeek. http://fifthelephant.in/2012 and http://hasgeek.com
Views: 12962 HasGeek TV
PyTorch Demystified, Why Did I Switch: Sherin Thomas
 
31:54
PyTorch entered into the realm of DL framework with the promise of being “Numpy on GPU”. The obvious failures of static graph implementation for certain use cases is increasing industry wide adoption of PyTorch. Dynamic Computation Graph being the backbone of PyTorch, comes with some perks. - Super intuitive - Shallow learning curve - Amazing community and discussion forum - Easy debugging It puts you explicitly in control of your compute, there’s no compiler trying to be clever and “help you” or speed your code up, but in reality mostly leading to a massive amount of debugging headaches. PyTorch should be the go-to framework for the problem with “dynamicity” in the core, especially for RNNs and highly dynamic Reinforcement Learning algorithms. My talk would be to convince the audience about the advantages of PyTorch. For what problems pytorch would be the best solution. I will have benchmarks with popular DL frameworks that review speed, performance, memory, and easiness. I’ll try to build on top of a comparison foundation with TF and NumPy so that the audience could correlate with their day to day research/development projects. The demo would be to portray the best cases where pytorch really shines over other DL frameworks. Note: Code I developed for the presentation would be available under MIT licence Slides and more info: https://anthillinside.talkfunnel.com/2017/20-pytorch-demystified-why-did-i-switch
Views: 14056 HasGeek TV
Michelangelo: Uber's machine learning platform - Achal Shah
 
44:57
Uber Engineering is committed to developing technologies that create seamless, impactful experiences for our customers. We are increasingly investing in Machine Learning to fulfill this vision. At Uber, our contribution to this space is Michelangelo, an internal ML-as-a-service platform that democratizes machine learning and makes scaling AI to meet the needs of the business as easy as requesting a ride. In this talk, I’ll go over some of Uber’s early challenges at applying ML at scale, and the context around which Michelangleo was born. We’ll also talk about what the Michelangelo system looks like, and some important components that aim to lower the bar on applying ML at Uber. Achal is a Sr. Software Engineer working on Michelangelo, and Deep Learning infrastructure
Views: 1558 HasGeek TV
Interview with Kalyansundaram, Media.net on necessary skills to become a DevOps engineer
 
09:42
We speak to Kalyan from Media.net on taking 1 AM calls, getting started with DevOps and skills to develop.
Views: 1391 HasGeek TV
Dr. Pramod Varma and Regunath Balasubramanian - Big Data at Aadhaar
 
50:30
Dr. Pramod Varma and Regunath Balasubramanian from the Aadhaar project describe the technical architecture of the world's largest biometric identity project at The Fifth Elephant 2012. The Fifth Elephant is a conference on the big data ecosystem by HasGeek. http://fifthelephant.in/2012 and http://hasgeek.com. Video and slides at http://hasgeek.tv/fifthelephant/2012-2/64-aadhaar-worlds-largest-biometric-identity-platform Slides at http://www.slideshare.net/regunathbalasubramanian/aadhaar-at-5thelephantv3
Views: 8880 HasGeek TV
Nikhil Ketkar - Are these the same pair of shoes? Matching retail products at scale
 
43:36
Product matching is the problem of resolving product entities across e-commerce sites. This involves a complex sequence of tasks which include - 1) automatic extraction of key information regions from raw HTML (for example, product titles, UPCs etc.) 2) categorising products into a unified taxonomy 3) semantic parsing of product titles and specifications 4) standardization of attributes such as brands, colours etc., 5) grouping products into clusters of matched products based on a similarity function or inferencing model. This is a challenging problem because unique and universally agreed upon identifiers are not always available and product details are noisy and often sparse. So we have to develop contextual understanding of product specifications, which are often expressed differently by retailers, merchants, aggregators etc. To scale the matching problem to half a billion products, we also need to prune and bucket effectively while achieving good recall. Matches need to be highly precise since customers may use them for sensitive tasks such as price comparison, competitive analysis and catalog enrichment. We employ an ensemble of online and offline algorithms and models to perform matching at scale for a large number of stores, categories and brands.
Views: 1831 HasGeek TV
Mohamed Imran K R - Scale anything with haproxy
 
16:47
In this talk, Mohamed Imran will: * Introduce haproxy as a webscale load-balancer capable of scaling any workloads * Demonstrate haproxy as a solution to scaling webservers, webservices, DB and discuss just about scaling any specific technology the audience wants to discuss * Discuss load-balancing strategies with an example * Discuss specific architecure solutions for magento, wordpress, drupal and elucidate how even a structureless/frameworkless app can be scaled * Introduce inotify based sync mechanism as an alternative to shared storage for use in load-balanced webservers * Introduce to security features of haproxy like rate-limiting and blackholing techniques to avoid webservers from being overwhelmed by specific attacks (should not be taken as DDoS protection schemes) * Demonstrate any audience query on load-balancing with an example (time permitting, we can do a full deployment)
Views: 4601 HasGeek TV
Introduction to SaltStack
 
02:01
In this talk, Mike Place will cover the basics of using SaltStack to configure servers and workstations through Salt's powerful remote execution and configuration management engine. Mike will provide overviews of many of the core concepts and then provide demonstrations of each concept so that each attendee can immediately begin to use Salt to solve everyday problems in their IT environments.
Views: 1225 HasGeek TV
Sunil Pai - Amplify your stack - Deux
 
32:59
Things have changed! Front end developers now mess with servers! Immutable data has gone from being a fad to a Very Good Thing™! And we’re still not writing unit tests! Time for another roundup of what frameworks and libraries and concepts (oh my) are the rage, which ones are actually useful, and what you should be investing in for your stack and your mind.
Views: 747 HasGeek TV
Developing Augmented Reality Android Application
 
31:09
Samarth Shah covers three main points. - Understanding Unity3d game engine - Understanding Vuforia SDK - Getting started with Unity3d (Eclipse IDE) and Vuforia SDK
Views: 47582 HasGeek TV
Scaling to 2000 request per second with MongoDB
 
16:48
In this video, Suchit Puri shares his experience with developing , scaling to 2000 req/sec and performance testing the website of one UK's biggest media houses. He covers: 1) Importance of "Schema Design" when scaling Mongo. 2) Mongo sharding. 3) Locking mechanisms in Mongo. 4) Replication mechanisms in a Mongo shard. 5) Performance testing the application. 6) Good sides of Mongo, mainly the flexibility it provides to your application. 7) Problems faced with migration.
Views: 4759 HasGeek TV
Cartonama 2012 - Geohash System and "Nearby" searches
 
20:08
Cartonama (http://cartonama.com/2012) is a conference on geospatial data, technology, apps and the community.
Views: 1502 HasGeek TV
Node.js Hacknight Pune - Participant projects.
 
02:54
Participants talking about the projects they are working on during the node.js hacknight (https://hacknight.in/jsfoo/2013-nodejs) that was organized at Pune. http://jsfoo.in/ and http://hasgeek.com/
Views: 1450 HasGeek TV
Amit Kapoor - Visualising Multi Dimensional Data
 
34:06
Even though exploring data visually is an integral part of the data analytic pipeline, we struggle to visually explore data once the number of dimensions go beyond three. This talk will focus on showcasing techniques to visually explore multi dimensional data p 3. The aim would be show examples of each of following techniques, potentially using one exemplar dataset. Standard 2D/3D Approaches Aesthetics e.g. Color, Size, Shape Small Multiples e.g. Trellis / Facets Matrices Views e.g. SPLOMs 3D Scatterplot Geometric Transformation Approaches Alternate Coordinates e.g. Parallel, Star Projections e.g. Dimensionality Reduction Tablelens Glyph based Approaches Star glyphs Stick Figures Pixel based Approaches Pixel bar charts Space filling curves Stacked based Approaches Dimensional Stacking Hierarchical Axis Treemaps The talk will also explore the role of interaction approaches to enhance our ability to visually explore the multi dimensional data. Interactive Approaches Navigation - Pan, Zoom, Scale, Rotate Selection & Annotation Filtering - Highlighting, Brushing and Linking Layering Dynamic Queries
Views: 2971 HasGeek TV
Jacob Appelbaum - Surveillance and Privacy
 
02:05:42
Jacob Appelbaum speaks about surveillance and privacy at a geekup on May 28, 2013, in Bangalore, India. The geekup was organized by HasGeek and Tactical Technology Collective in association with null - The Open Security Community. http://geekup.in/2013/ioerror Foreword by Bhairav Acharya, legal researcher working on the draft Privacy (Protection) Act of 2013 in India. We lost a few seconds of Bhairav speaking, but the bulk of his foreword and the entirety of Jake's talk are present here. Video by HasGeek TV - http://hasgeek.tv/geekup/ioerror/567-jacob-appelbaum-surveillance-and-privacy
Views: 12054 HasGeek TV
`Rxify`— a simple spell for complex RxJava operators - Garima Jain, Fueled
 
44:02
RxJava operators are very useful but a little hard to learn at first. The task of selecting the appropriate operator for a particular scenario. The purpose of this talk will be to make the selection task easier. Giving examples of day-to-day scenarios and explanation of how to apply them. The rough list of the operators which I will talk about is as follows concat() debounce() defer() take() repeat() zip() skip() join() filter() retry() buffer() Others I will start with the definiton of the operators, explain what it does and then list a scenario in which it can be applied. Will then show a simple code snippet of how to appy those operators. A gist of the talk that I will be presenting can be found on related blog posts on medium. https://droidconin.talkfunnel.com/2016/94-rxify-a-simple-spell-for-complex-rxjava-operators
Views: 1508 HasGeek TV
Srihari Sriraman - Of the building of a Postgres cluster
 
43:17
We engineered a Postgres database cluster last year. It was a lot of learning and a lot of fun! This talk is about the failure scenarios we designed for, the times when the designed system failed, and what we learnt from them. https://rootconf.talkfunnel.com/2016/15-of-the-building-of-a-postgres-cluster
Views: 2834 HasGeek TV
Nishant Bangarwa – Scalable Realtime Analytics using Druid
 
45:33
Traditional SaaS solutions based on hadoop datastore Hive/Hbase or classical RDBMS work well for storing data, although they are not optimized for ingesting data and making it immediately available for interactive ad-hoc low latency queries at a very high scale. Long query latencies make these solutions suboptimal choices to power interactive applications. This talk will introduce Druid as a complementing solution for scalable real-time ingestion and analytics. Druid is an open source distributed data warehouse, designed to support OLAP-like queries and is used in production at numerous companies. It was inspired by Google’s Dremel, PowerDrill and search framework. This talk will cover druid architecture, its storage internals and the common use cases druid is a good fit for.
Views: 2354 HasGeek TV
Devashish Shankar - Deep Learning for Natural Language Processing
 
44:43
Much of the Text Mining needed in real-life boils down to Text Classification: be it prioritising e-mails received by Customer Care, categorising Tweets aired towards an Organisation, measuring impact of Promotions in Social Media, and (Aspect based) Sentiment Analysis of Reviews. These techniques can not only help gauge the customer’s feedback, but also can help in providing users a better experience. Traditional solutions focused on heavy domain-specific Feature Engineering, and thats exactly where Deep Learning sounds promising! We will depict our foray into Deep Learning with these classes of Applications in mind. Specifically, we will describe how we tamed Deep Convolutional Neural Network, most commonly applied to Computer Vision, to help classify (short) texts, attaining near-state-of-the-art results on several SemEval tasks consistently, and a few tasks of importance to Flipkart. In this talk, we plan to cover the following: Basics of Deep Learning as applied to NLP: Word Embeddings and its compositions a la Recursive Neural Networks, Convolutional Neural Networks, and Recurrent Neural Networks. New Experimental results on an array of SemEval / Flipkart’s internal tasks: e.g. Tweet Classification and Sentiment Analysis. (As an example we achieved 95% accuracy in binary sentiment classification task on our datasets - up from 85% by statistical models) Share some of the learnings we have had while deploying these in Flipkart! Here is a mindmap explaining the flow of content and key takeawys for the audience: https://atlas.mindmup.com/2015/06/4cbcef50fa6901327cdf06dfaff79cf0/deep_learning_for_natural_language_proce/index.html We have decided to open source the code for this talk as a toolkit. https://github.com/flipkart-incubator/optimus Feel free to use it to train your own classifiers, and contribute!
Views: 11339 HasGeek TV
Gene Ekster – The Alternative Data revolution on Wall St
 
01:03:51
This talk will focus on the role that non-traditional data research, known as alternative data, is beginning to play across the investment community. We will address how datasets such as point of sale transactions, web site usage, municipality records, social media data and similar information are being utilized by traditional long-short funds, quantitative hedge funds and also mutual funds. Topics covered will include aspects of the developing alternative data ecosystem including: * Alternative data R&D process flow * Computing infrastructure and the technology stack * Research & analytics providers * Technical solutions to common issues found in alt. data * Best practices We’re going to walk through a few examples of how noisy, unstructured data become an investable signal using tools such as text mining and machine learning. The aim is to introduce the audience to the process of how hedge fund portfolio managers and sell-side research analysts are systematically generating returns by leveraging unique primary (bots / scrapers, channel checks) and third party datasets (including data brokers). This includes sourcing, compliance, scrubbing out PII, alpha generation related to revenue estimates and approaches to balance the secret sauce with product transparency. Finally, we’ll ponder the future of alternative data in finance and touch on how companies in the data space can best take advantage of this growing trend. Gene Ekster was previously head of R&D at Point72 Asset Management (formerly SAC Capital), a Director of Data Product at 1010Data and a Senior Analyst at Majestic Research (now ITG Investment Research). Currently, Gene works with asset management firms and data providers in a consulting capacity to help integrate alternative data into the investment process. He can be reached via LinkedIn (https://www.linkedin.com/in/geneekster). This talk was recorded at The Fifth Elephant 2016, India's premier data analytics conference.
Views: 2169 HasGeek TV
Anand S - Beautiful visualizations with d3.js
 
01:00:55
"D3.js is a JavaScript library for manipulating documents based on data. D3 helps you bring data to life using HTML, SVG and CSS." For more examples, see http://d3js.org/ This workshop will walk through the concepts of creating visualisations with D3 and dive into specific examples that you'll be working on. This talk was recorded at JSFoo 2012, a JavaScript conference by HasGeek. http://jsfoo.in/2012 and http://hasgeek.com
Views: 30939 HasGeek TV
Mike Place - Introduction to SaltStack
 
58:25
In this talk, we will cover the basics of using SaltStack to configure servers and workstations through Salt’s powerful remote execution and configuration management engine. We will provide overviews of many of the core concepts and then provide demonstrations of each concept so that each attendee can immediately begin to use Salt to solve everyday problems in their IT environments. Mike Place is the Sr. Platform Developer on the core engineering team at SaltStack. He works primarily on SaltStack core internals, with a focus on performance and optimization. He has twenty years of experience in UNIX and Linux systems engineering across a wide variety of environments.
Views: 20353 HasGeek TV
Satnam Singh - HawkEye: A Real Time Anomaly Detection System
 
19:42
HawkEye is a real-time anomaly detection framework for detecting anomalies in IT infrastructure data e.g. CPU, memory, response time of a data center machine. The framework uses a combination of anomaly detection techniques to detect local and global anomalies. The system detects several types of local anomalies using different anomaly detection techniques over a sliding window. We used rigorous data experiments to perform data preprocessing and select hyperparameters. We have employed statistics-based techniques to detect local point anomalies and statistical detection theory (Page’s Test) to detect local contextual anomalies. Time-series models are employed to detect seasonality in the data and detect global anomalies. In this talk, I will showcase results of the HawkEye system on real-data.
Views: 1993 HasGeek TV
React Native in production - Mario Stallone, Myntra
 
45:54
React Native for Android was announced about a year ago, and there has been a lot of speculation over using it with large apps. We’ve been using React Native in our apps that reach millions of users, for quite sometime now, and we wanted to share our experience with you, so that you too can take that leap of faith and reap the benefits that we did. Benefits that we reaped - Increased developer efficiency - Faster release cycles - More developers chipping in. All front end devs, can now do mobile… Yeah! - Instant Refresh - Simpler Code and… - Share code between Android and iOS - Near instant release cycles In the talk I will focus on our journey in taking React Native to production and talk about the benefits we reaped, some pitfalls/caveats that we faced, etc. https://droidconin.talkfunnel.com/2016/87-react-native-in-production
Views: 4395 HasGeek TV
Pranay Dubey - JavaScript for Competitive Programming
 
19:41
Competitive programming requires lots of Data Structures / Algorithm knowledge to be able to perform efficiently. Though this knowledge is helpful, there are certain optimizations that can be done in the language(JavaScript) itself that would allow the program to be executed faster. These optimizations can be done efficiently if we get an idea of ‘behind-the-scene’ execution of our program. JavaScript lacks in two main areas which(as of now) makes it unfavourable for competitive programming: 1.Speed - Includes programs like calculating 25000th prime number which is around 5x times slower in comparison to c,c++ using the same algorithm. 2.Memory - Includes programs like calculating factorial of a big number(say 1000) while in JavaScript you get ‘infinity’ for factorial of any number above 170 because it exceeds the number limit. The talk will be focused on discussing these limitations and ways to overcome them, thus making JavaScript a suitable choice for Competitive programming. https://jsfoo.talkfunnel.com/2015/16-javascript-for-competitive-programming
Views: 1535 HasGeek TV
Suraj Srinivas - Making Deep Neural Networks Smaller and Faster
 
17:47
Deep neural networks with millions of parameters are at the heart of many state of the art machine learning models today. However, is has been shown that models with much smaller number of parameters can also perform just as well. A smaller model has the advantage of being faster to evaluate and easier to store - both of which are crucial for real-time and embedded / mobile applications. In this talk, I intend to provide a brief overview of such model compression techniques. Using these techniques, it is possible to compress neural networks by as much as 10x and speed up inference by 3-4x.
Views: 893 HasGeek TV
Arjun Jain - Joint Training of a CNN and a Graphical Model for Human Pose Estimation
 
12:34
We propose a new hybrid architecture that consists of a deep Convolutional Network and a Markov Random Field. We show how this architecture is successfully applied to the challenging problem of articulated human pose estimation in monocular images. The architecture can exploit structural domain constraints such as geometric relationships between body joint locations. We show that joint training of these two model paradigms improves performance and allows us to significantly outperform existing state-of-the-art techniques.
Views: 601 HasGeek TV
Adapting bandit algorithms to optimise user experience at Practo: Santosh GSK
 
18:00
The art of trading between exploiting the best arm versus exploring for further knowledge of other arms has long been studied as Bandit Algorithms in various fields of clinical trials, designing financial portfolios, etc. Recently, in website optimization, these algorithms have been used for optimizing click through rates and performing A/B testing. However, these algorithms have the potential to be applied in several other contexts where we need to optimize for a reward by exploring possible arms. In this talk, Santosh would be presenting an approach based on Contextual Multi-armed Bandit algorithms for achieving a better tradeoff between user’s expectation of faster replies and doctor’s burnout rate on a QnA platform like Practo Consult.
Views: 437 HasGeek TV
Meta Refresh Previews - Sunil Pai
 
03:16
Sunil Pai introduces his talk on front-end build process at Meta Refresh 2012. http://metarefresh.in, 21st April 2012, Bangalore
Views: 1309 HasGeek TV
Abhishek Narain - Building 3D Games using Babylon.js
 
41:26
WebGL is integrated into all the web standards of the browser allowing GPU accelerated usage of physics and image processing and effects as part of the web page canvas. WebGL elements can be mixed with other HTML elements and composited with other parts of the page or page background. Babylon.js is a 3D engine based on webgl and javascript. In this talk we would look at the capabilities of Babylon.js and its advantages. Post which we will build a Getting started game! https://jsfoo.talkfunnel.com/2015/27-building-3d-games-using-babylon-js
Views: 14395 HasGeek TV
Joydeep Sen Sarma - Messaging architecture at Facebook
 
30:26
Joydeep Sen Sarma, an ex-Facebook employee who participated in the design of the Facebook messaging system, describes its architecture at The Fifth Elephant 2012. The Fifth Elephant is a conference on the big data ecosystem by HasGeek. http://fifthelephant.in/2012 and http://hasgeek.com.
Views: 7471 HasGeek TV
Anand Chitipothu - 10 reason why you should prefer PostgreSQL to MySQL
 
43:43
MySQL is the most popular database server, but popular doesn’t always mean the best. PostgreSQL on the other hand is less popupar, but claims to be the most advanced database server. In this talk I plan to do a in-depth comparision of MySQL and PostgreSQL on various fronts. I’m an independent software consultant and trainer. I’ve been using PostgreSQL for long time. During my work at Internet Archive, I’ve migrated their database server from MySQL to PostgreSQL and learnt lot of things about both of them during that.
Views: 5084 HasGeek TV
Suchakrapani Datt - Kernel and userspace tracing with LTTng and friends
 
36:33
Tracing is an important tool for diagnosing performance issues in critical applications (kernel/userspace). Sometimes debugging is just not enough as it may distort the application's flow itself. This talk shall discuss the most important features of the commonly used tracing tools (LTTng, SystemTap, Dtrace) and some surprisingly cool features they have (such as fast tracing from inside GDB)
Views: 903 HasGeek TV
The road to styled components, and the road ahead - Glen Maddern, Front End Center
 
46:59
Building user interfaces on the web is hard, because the web, and thus CSS, was inherently made for documents. Because interactive UIs fundamentally differ from flat documents, we’ve seen a mindset shift towards building component-based systems. That progression has lead to a rise in options for including your CSS in your JavaScript code, much the same way that libraries like React have done for your HTML. But what can we learn from the progress of the past, and what do we choose to include in a new library? This talk will cover how the design goals of styled-components have helped it become the most popular alternative for CSS-in-JS. But more broadly than that, it will look at the progression from documents, through components, to an exciting new possibility of an inclusive, unified workflow for UI on the web and elsewhere.
Views: 1008 HasGeek TV
Muktabh Mayank - Making a contextual recommendation engine
 
14:42
Making a Recommendation Engine at ParallelDots a. Why normal full-text search will not work: The problem of incorrect tagging and slow search queries. b. ParallelDots’ MVP with Topic Models: Issues with accuracy and scaling. c. Decision to use Deep Learning and aims of the new architecture (Not enough funds for distributed system, search related posts from millions of documents in reasonable time) Basics of Deep Learning a. Deep Neural Networks b. Types of Deep Neural Networks. Convolutional, DBNs, Recurrent and Recursive. How do they differ in structure, types of neurons and training. c. Backpropogation and its variants d. Features of various Deep Learning libraries in Python. Deep Learning in NLP a. Solving problem of high-dimensionality using word embeddings. b. Common approaches to word embedding. c. Modelling language as a series of characters using Recurrent Neural Networks . c. Models we use : Named Entity Recognition with Neural Nets d. Models we use: Combining word embeddings using heuristics and recursive neural networks. Search Engine a. Using Search Data Structures to convert search related posts operations from O(n) to O(log(n)) b. Space Partitioning Trees : Search for nearest Neighbours. Examples of such trees: KD-Tree / Ball Tree / VP Tree c. Why we chose VP Tree ? What libraries to use to code up in Python ? d. Parallelization. Data Parallel Python’s multiprocessing parallelization not the best, working towards a shared memory parallel version. Scaling up system a. Hacks to scale up recommendations. b. Using golang’s channels to unique requests.
Views: 2252 HasGeek TV
[Flash Talk] Jaseem Abid - Emacs Meetup
 
05:21
Jaseem introduces Emacs and the Emacs Meetup to the Rootconf crowd.
Views: 479 HasGeek TV
Kumar Rangarajan - Android battery consumption: the what & how
 
30:46
Kumar Rangarajan from LIttle Eye Labs shares tools and techniques for measuring, managing and optimizing battery power consumption in Android apps, and then introduces his startup's battery usage profiling tool for developers. Recorded at Droidcon India 2012, an Android developer conference by HasGeek. http://droidcon.in/2012 and http://hasgeek.com
Views: 992 HasGeek TV

The tell tale heart annotated bibliography
What is a cover letter supposed to say
Cyberwire newsletter formats
Help with writing resume and cover letter
Sample cover letter early childhood teacher