Software Engineer
I work at the intersection of backend systems and applied AI, building scalable, resilient services and deploying intelligent models in production — from cloud-scale Tier-1 systems to resource-constrained edge devices.
Lessons learned building a reliable task queue from scratch with Go, Redis, and gRPC.
A practical guide to building responsive terminal interfaces using Rust and the ratatui crate.
A deep dive into Conflict-free Replicated Data Types and how they power real-time collaborative editing.
Optimized LLMs and vLLMs using QAT and PTQ, reducing model size up to 4× and cutting inference latency by 30–40%. Enabled deployment of generative AI models on resource-constrained edge chips while maintaining ~95% accuracy.
Built a VRAE-based pipeline to embed UWB time-series data for classification and autofocus distance prediction, shipped in Samsung S25. Prototyped DeepCluster-based unsupervised learning to extract structure from unlabelled UWB data.
Worked on asset management for Amazon Detail Page. Strengthened CI/CD pipelines by adding automated latency regression checks through real-time response marker parsing, improving reliability at scale.

A full-stack rental platform for two-wheelers with customer, owner, and manager roles. Implemented payments, asset storage, and role-based access control.

A programming language interpreter built in Go with a REPL, supporting variables, expressions, statements, and functions.