01 — HELLO
Founding Engineer @ Soxton AI
02 — ABOUT
Computer vision, deep learning and VR — built to run on real hardware.
I’m a Founding Engineer at Soxton AI in New York. I graduated from New York University with an M.S. in Computer Science (2026, GPA 3.92/4.0), where I was a Teaching Assistant for Deep Learning, a Numerical Computing tutor at Courant and a Research Assistant in AI & neuroscience. My technical paper on dichoptic foveation for VR rendering optimization is published at ACM SIGGRAPH 2026. Before NYU I earned a B.Tech in Data Science Engineering from Manipal University Jaipur with distinction (8.90/10) and the Dean’s List Excellence in Academics Award.
My research spans facial-morphing detection on the FISWG 19-feature framework with Prof. Banerjee and vision models for behavioral neuroscience with Prof. Froemke, with a focus on lightweight AI — like distilling 72B models into efficient 7B architectures. At Hewlett Packard Enterprise, ONGC and OnFees I built server-management systems that cut crashes by 30%, predictive-maintenance models that cut downtime by 25%, and real-time analytics platforms. I’m a three-time hackathon winner, including the GIDE Hackathon at a16z NYC Tech Week and the NVIDIA × Antler × Acer NYC Spark Hack, where my team deployed a real-time public-safety vision system entirely on local hardware. I work in Python, C++, Java, JavaScript and Dart, with TensorFlow, PyTorch, Hugging Face, AWS, Docker and Kubernetes.
03 — RESEARCH
Prof. S. Banerjee, NYU
Prof. Robert Froemke, NYU Langone Health
Prof. Qi Sun, NYU Immersive Computing Lab
04 — PUBLICATION
VR rendering optimization
Designed and validated a dichoptic foveation rendering technique through 2AFC psychophysics studies with 50 participants, achieving a 30–40% reduction in rendering compute while maintaining perceptual equivalence. Built in Unity (C#) with a real-time dichoptic rendering pipeline running at 60+ FPS across diverse VR hardware.
05 — EXPERIENCE
Soxton AI · Full-time
Hewlett Packard Enterprise (HPE)
Oil and Natural Gas Corporation (ONGC)
OnFees
M.S. in Computer Science
Coursework Big Data, Artificial Intelligence, Deep Learning, Open Source Development, Information Visualization, Design & Analysis of Algorithms
Unofficial transcript for New York University (PDF, opens in a new tab)B.Tech in Data Science Engineering
Coursework Data Structures, OOPS, Computer Networks, OS, Machine Learning, NLP, Artificial Intelligence, Big Data Analytics
Official transcript for Manipal University Jaipur (PDF, opens in a new tab)06 — PROJECTS
01
Enterprise-grade DNS filtering that blocks malware, phishing and ads with real-time threat intelligence. A caching Go resolver answers in under 15 ms, behind a Node.js API gateway and a React dashboard.
02
Cross-platform mobile app with AI-powered food recognition (>90% accuracy) and multimodal logging — camera, barcode, voice and text. Firebase backend, Clean Architecture and the BLoC pattern.
03
A real-time multi-user cloud IDE where every developer brings their own AI agent. An orchestrator AI prevents conflicts, manages file locks and routes agents to non-overlapping work.
04
Person of Interest DECK/01 — a local-inference tactical HUD for NYC predictive surveillance. Processes live feeds from 900+ traffic cameras entirely on-device.
05
Scrapes job listings, scores fit, generates tailored one-page LaTeX resumes and emails the top matches — all while you sleep. End-to-end automation in Python.
06
Combines NYC public datasets — weather, taxi pickups, subway ridership, 311 requests and events — into interactive visualizations that show how urban factors correlate across boroughs over time.
07
Free, open-source alternative to ShareMouse: one mouse, keyboard, clipboard and file transfer shared between a Mac and a Windows PC over the LAN. TLS-encrypted with fingerprint verification, no accounts or cloud relay.
07 — HACKATHONS & LEADERSHIP
Winner · Perkins Coie LLP, New York · June 2026
Won at an elite 50-person, in-person hackathon hosted at a16z NYC Tech Week. Built a core solution in GIDE powered by a locally run Qwen Coder 4B model and demoed it live to judges, competing on technical merit, inventive use of GIDE, effectiveness and creativity.
1st place · NVIDIA × Antler × Acer · 2026
Built Person of Interest: an end-to-end ML system that analyzes live feeds from NYC’s 900+ public traffic cameras with a vision-language model (Qwen3-VL-30B) to detect dangerous activity. Fine-tuned on a custom annotated fight-scenario dataset and deployed fully locally on an Acer Veriton GN100 (NVIDIA GB10, 128 GB unified memory). Won an Acer Veriton GN100 workstation and an invitation to present at the NVIDIA AI LinkedIn Live showcase.
1st place · Betaworks, New York · May 2026
Won the $1,000 cash prize for the best project built with chat.dev, a platform for programming on the go from your phone. The event opened with a 5K run, and the winners presented at Betaworks.
Mentor · IC Hack 2023
Mentored 10+ participants, guiding app development and sharing technical expertise in mobile application development and software engineering best practices.
08 — SKILLS
09 — CONTACT
Research, AI engineering, or a good idea worth building — I’m always happy to connect. Say hello.