About
I am a 2nd-year B.Tech CSE student at NIT Mizoram. Upstream contributor to OpenCV (Intel-led). Specializing in IoT Systems, Scalable Backends, and Applied ML. Proficient in architecting full-stack applications with 3NF-normalized databases, real-time telemetry pipelines, and predictive ML models. Multi-time coding contest winner. Team Head of the Coding & AI Club.
JPMorgan Chase & Co. (Forage) — Software Engineering Virtual Experience (Mar 2026): Engineered a real-time data visualizer using Apache Kafka and H2 DB, achieving sub-10ms response latency for high-frequency live data ingestion flows.
Google Cloud Rapid Agent Hackathon — AI Developer (May 2026): Developing a "Crisis Response Agent" focusing on multi-step AI agents capable of taking real-world actions, utilizing GCP and RAG for dynamic data retrieval.
My primary interest is Edge AI and Model Quantization. I independently connected bit-width representation from Computer Architecture lectures to weight compression in neural networks — which led me to study Binary Neural Networks, 1-bit LLMs (BitNet, BitNet b1.58), and on-device inference pipelines before formally studying machine learning.
I am currently developing a concept for a locally-running AI scam call detector for low-end Indian smartphones — zero cloud dependency, no data sharing, using quantized speech models on-device.
In September 2025, I diagnosed a floating-point precision discrepancy in OpenCV caused by MinGW's default FPU extended-precision mode on Windows. I traced the root cause through the production C++ codebase, applied a targeted patch, and submitted a Pull Request to the upstream repository.
I serve as Team Head — Website & Social Media at the Coding & AI Club, NIT Mizoram. I lead the official club website, run weekly DSA and Git workshops for 50+ students, and founded a mentorship program where beginners reach medium-level competitive programming within one semester.
Skills