DevNetwork hosts over 30,000+ developer & technical professionals per year at 10+ global conferences & events, including DeveloperWeek, API World, the DeveloperWeek Global Series, AI DevWorld, and more.Helping shape each conference's content, vote on the annual DEVIES Awards, and judging the hackathons.

Working as Technical Lead Owner in the Video Infra team, responsible for the end-to-end video infrastructure of all Meta products (FaceBook, Instagram, Whatsapp, Messenger, Oculus). My work resulted in a 3% revenue increase. Our infra process 1+ Billion videos/day. • Designed compute-efficient scalable HDR architecture that resulted in a 50% reduction in compute.Increased user watch time by more than 3%, resulting in Ad revenue growth by enabling HDR videos.• Lead the cross-functional teams for client side design for video upload and playback ( iOS, Android, Browser, Oculus teams).• Devised bandwidth-efficient HDR ABR ladder in collaboration with multiple research Labs ( UT Austin, Nantes France, AGH Poland, RISE Sweden), which resulted in 20% reduction in bandwidth consumption for HDR delivery.• Manage summer interns for small high-impact projects• Industry-wide alignment with Apple, Google (Android), Samsung, Xiaomi to align on the HDR roadmap.• Involved in the design of Video Accelerator by providing Resampler/Tone Mapping, and reference implementation.• Work on integrating video pipeline for remote rendering, graphics, and remote sensor access in Oculus Developer Tool Infrastructure.

Worked in the GeForce Now team of Nvidia. This group is responsible for streaming 4K video games from Cloud and from GPU-enabled Gaming Desktop at home to Tegra-powered Shield devices and Win/Mac clients. My key contribution so far involves the following:• Lead GFN-PC reliability Squad.• Root-caused low throughput problem of the reliable UDP (ENet) channel, (increased throughput from 1.5mbps to 250mbps).• Make the ENet library thread-safe and stabilize it for production. • Increased GFN-PC yield by ~3% by successfully replacing the GameStream control channel (Mouse input, QoS etc.) from TCP to Reliable UDP.• Implemented and shipped successfully NvExtendedCommand/Event channel for secure transmission of commands from Client to Server (e.g. Serenity)• Found, root-cause of slow control channel and its fix.• Lead transition of old Rtsp server to Rtsp-over-POCO.• Lead Secure Rtsp effort.• Performed discovery of Gaming Desktop on the Tegra-powered device using Zero Configuration protocols (DNS, mDNS, DNS-SD, DHCP, and link-local addressing). The connectivity is enhanced from ~80% to ~95%.• Enabling multimedia traffic priority by setting DSCP using Windows QoS API. It helped to reduce the latency of video pipelines in LAN scenarios. In a WAN scenario, routers normally do not pass DSCP settings.

Emerging Products group is responsible for developing products based on latest technologies like High Dynamic Range (HDR), Intelligent tone mapping and metadata processing technology. Driving a group of 5 engineers, while closely working with Product Managers and team-members, my development scope spans the following areas:• Enhanced the development speed of whole team by several times by designing multi-VM Vagrant environment for whole team.• Reduced deployment time from 90 minutes to few seconds by converting software modules to PIP packages and RPMs.• Create low-cost secure PIP and RPM repositories for deployment.• UI test automation with Selenium framework.• Development with AWS EC2/S3 APIs in Boto python package.• Enhanced SAAS-architecture for cloud deployment to reduce cost by 300%.• Developed an auto-scaling architecture which works both in EC2 and on-premise data centers.Deployment:• AWS Console• AWS CLI, S3CMD, Python BOTO Package• Vagrant• Jenkins (Continuous Integration)Multimedia Pipeline:• Tuttle OpenFX pipeline• FFMpeg, Mediainfo, Sam command-line toolsWork-flow tools:• Atlassian’s JIRA tool for agile development.• GitHub and SourceTree for source-control. SW Quality by ensuring:• Test Driven Development (TDD).• Jenkins jobs are configured with Code Coverage bar in place for all modules and commits degrading the coverage are rejected• For python, pep8/pylint style checkers are utilized to ensure proper coding practices.

EagleyeMed was developing a tele-medecine platform. Major tasks accomplished are:• 50 mSec end-to-end latency was achieved for transmission of medical imaging.• US Patent is granted for diagnostic quality compression at low bitrate.• Avoid $50k/developer cost by mocking the Medical imaging device for development.• Implemented the whole imaging and video processing pipelines (capturing, pre processing, compression, streaming, post processing) for different platforms (embedded, windows) from scratch.• Avoid port-opening by implementing UDP Hole Punching.• HIPAA-compliant secure transmission was achieved by DTLS (Datagram Transport Layer Security).• Developed in-house algorithms for handling lost & out of order packets.• Created a detailed test document for complete functional and stress testing of platform.Tools:C/C++, STLH.26X/MpegXUDP/TCP Streaming, STUNMySql, PhpMyAdmin, WireSharkJira, Github

• Our video encoder was used to stream SuperBowl 2014 to more than 528,000 viewers simultaneously online, which was a world record for online viewing of any event. • Added support for new video formats in transcoding framework of TV Everywhere platform.• Support for Telecine video content was implemented for VOD content.• Fixed bugs in framework and inside video codecs for different video formats.Tools:C/C++, Lua, Shell scriptingVideo Codecs (MpegX, H.26x, Prores, DNxHD)Fedora, CentOSSVN, RedminePHP, MySQL

• Worked as independent architect for media encoding to support Cloud Gaming. • Provided technology leadership and vision on video encoding implementation for TI SOCs DM365, 368 and DM8148 (Centaurus) platforms• Design and implemented the following projects: 1. Adaptive bitrate control for dynamic nature of Internet. 2. Low latency video coding. 3. Dynamic rate control adaptive to available bandwidth. 4. Adaptive Intra Refresh for error resilience.Tools: C/C++, MatLabH.264,FFMpeg, Openmax ILembedded linuxGIT, Jira, Jenkins,

During this period, I was working in Labsalongwith my Master, PhD and PostDoc. In my PostDoc (IRCCyN Labs, Nantes France) during Jan-Sep 2011 , I worked on QOE (quality of experience) for scalable video codecs and design of new watermarking algorithms. I also worked on drift free bit stream watermarking of H.264/AVC.During my PhD (LIRMM Labs Montpellier France, Jan'08-Dec'11) I designed world first real-time format complaint video encryption system. My IEEE TCSVT paper has been cited more than 50 times. I worked on the real-time protection of scalable video codecs using selective encryption, watermarking and active fingerprinting. For encryption, I have worked on CAVLC and CABAC of H.264/AVC and C2DVLC of AVS. For watermarking, I worked on IT of H.264/AVC.During my Master, I was working (LIRIS Labs Lyon France, Sep'06-Dec'07) I worked on use of Independent Component Analysis (ICA) for image and video compression. DCT transform basis were replaced by over-complete ICA basis. Transform was performed using Matching pursuit. The designed framework was found to be useful for low bitrate image and video compression.

• Worked for Research, Development, Optimization & Testing of OEM/ ODM Products based on Philips VLIW Processor TriMedia e.g. DVR, Video Surveillance, Video Telephone.• Involved in algorithm development, pre & post processing, memory management and system level issues of Video Codecs. • Involved in functional and stress testing of Video Codecs.Tools:C/C++, MatLabVideo Codecs (MpegX, H.26x)pSoS RTOSSource Control Management Tools (CVS) & Makefile