B.Tech | M.S - Computer Science
I build loosely coupled systems that don't break your sound sleep at 3 AM & currently leading the engineering team @ QuickSlot Health Inc, where I engineer a HIPAA-compliant healthcare platform for a clinical network across the United States. And still find time to obsess over lock-free data structures.
Leading QS Meridian — a HIPAA-compliant clinical documentation platform. Built the entire AWS backend (S3, Lambda, DynamoDB, KMS, Cognito). Drove 300% product growth and reduced infrastructure costs by 78% with a custom chunk-level validation algorithm.
Contributed a math.MaxInt64 function to the Hugo open-source repo in Go. Conducted PR complexity reviews and rubric evaluations for AI model trajectories across Project Ivy and Helix.
Built a multi-agent RAG system (Claude API, FastAPI) that hit 100% answer accuracy at 2ms latency across 100+ concurrent requests - outperforming DeepSeek R1 by 51%."
Migrated monolith to Kafka Streams microservices. Built observability pipelines with Splunk, Grafana, and Prometheus. Designed a C++ priority-queue scheduler for real-time workload distribution.
High-performance single-producer, multi-consumer queue in C++ for low-latency message distribution where every microsecond counts.
Real-time messaging infrastructure with fault-tolerant delivery and scalable concurrent connection handling.
This project involves developing a comprehensive data-driven system to evaluate, analyze, and re-rank recipes
CPP tool / extension for detecting and preventing resource leaks, making it easier to identify and fix memory management issues.
River network extraction is crucial to keep track of the water resources. Various methods have been implemented in times series to yield profound and incisive outputs and are still being developed and combined with predefined available methods. We have carried out a structured survey on these methods and have presented them with their outputs.
Surveyed segmentation and deep learning approaches for extracting river networks from satellite imagery, developing a UNet model that achieved an 80.98% dice score on Kaggle and Google Earth Engine datasets. Techniques generalize to broader semantic segmentation detection tasks.