May 2022 – Aug 2022 · 3 mos · Seattle, Washington, United States · On-site

• Built and demonstrated PyTorch-based components of synthetic data generation methods for Amazon's Just Walk Out (JWO) computer vision pipelines: 3D pose estimation and human body segmentation.
• Designed deep learning pipelines for pose estimation and novel view synthesis using techniques like: parametric human body models (SMPL), differentiable rendering, inverse kinematics, neural rendering fields (NeRFs).

Jan 2021 – Jul 2021 · 6 mos

• Worked on design, data capture, implementation and training of deep learning based models for hand gesture recognition on mobile devices.
• Achieved 90%+ gesture recognition accuracy on an in-house dataset by designing and training neural networks - TCNs, LSTMs, 1-D CNNs, etc.
• Achieved similar results on Android by porting the best networks among these.

Apr 2020 – Dec 2020 · 8 mos

• Researched, developed, and deployed solutions for Samsung Electronics devices. Solutions cover domains such as:
• Shadow rendering and animation engines for the AR Canvas app
• Automatic 3D model rig and animation synthesis
• Image processing
• 3D geometry processing for 3D printing

Jul 2018 – Mar 2020 · 1 yr 8 mos

• Handled development of graphics solutions for commercial release in Samsung mobile phones.
• Developed and deployed the Neon Glow pen effect in AR Doodle, an augmented reality drawing app, released in 1.3 million+ Galaxy Note 10 devices
• Won a Spot Award for this contribution
• Performed research for solutions to be deployed across Samsung Electronics devices, across domains in computer graphics such as:
• 3D geometry processing for 3D printing
• Cloth simulation and garment generation techniques
• Certified, Professional Level Software Competency, second highest level of software competency attainable at Samsung

Feb 2018 – May 2018 · 3 mos · Chennai Area, India

• Started a club of 15+ game development enthusiasts (artists, designers and programmers) to form a community of such like-minded people
• Conducted introductory game art and programming sessions for students
• Started a culture of mini-projects for learning purposes, like a small version of the indie shooter SuperHot

May 2017 – Jul 2017 · 2 mos · Bengaluru Area, India

• Developed a tool to find vulnerabilities in the Samsung Pay application, using the technique of fuzzing (feeding random inputs at the bit level to test code coverage of functions).

Apr 2017 – Jan 2018 · 9 mos · Chennai Area, India

• Headed a 6-member team tasked with designing and developing games or VR/AR software for publicity via the website of Shaastra, the annual technical festival of IIT Madras.
• Made the Survival 3018 time-travel shooting game as part of this role and deployed it to the Shaastra website; the game was played by 500+ users.

Apr 2016 – Jan 2017 · 9 mos · Chennai Area, India

• Headed a 6-member team tasked with designing and developing games or VR/AR software for publicity, for Shaastra, the annual technical festival of IIT Madras.
• Made VRoom, an Android local-multiplayer VR car-racing game for Google Cardboard, as part of this role.

Dec 2016 – Feb 2017 · 2 mos · Chennai Area, India

• I joined Crion as a trainee Unreal developer. I learnt Unreal Engine with Blueprints during that period and made contributions to some of their VR applications developed for the Oculus Rift.
• Areas I worked on include general Unreal scene setup, asset management, and Blueprint scripting. I learnt the basics of the Oculus SDK for Unreal as well as the native multiplayer support in Unreal.

Nov 2016 – Jan 2017 · 2 mos · Chennai Area, India

• This was an independent gaming studio initiative to develop refreshing and exciting content for mobile and VR gaming.
• I handled development of core features in Unity2D for the Android physics-based puzzle game Blubb as part of this.

Dec 2014 – Jul 2015 · 7 mos · Chennai Area, India

• I had contributed to the development of the Clozerr Android app for Tara Tech Solutions Pvt. Ltd., an IIT Madras incubated startup, to make an intelligent customer loyalty service for a variety of vendors. I went on to complete an internship there during May-July 2015.
• Areas I had worked on:
• 1. Beacon detection and beacon-based UX customization (during intern)
• 2. Geofence monitoring and user data collection and analytics (during intern)
• 3. UI and backend integration