Initializing
VM
Software Engineer
Work
Blue Horse Digital pulled me full stack with AI powered product features. Hexaware hardened me on APIs and LLM integration under real load. IBM taught me event driven backend at scale. Most recent first. Scroll through each chapter.
Feb 2026 to May 2026
01Full Stack Developer
When the backend sprints, the UI has to keep up
Full stack work on an event discovery product. Data moved fast on the backend, so the frontend could not afford to lag behind. React on the surface, GraphQL and RxJS wiring everything together underneath. Also shipped AI powered discovery: semantic search and real time recommendations fed by LLM and embedding pipelines, with the same obsession over latency and reliability as every other feature.
Dec 2024 to Dec 2025
02Backend Engineer
APIs that breathe under pressure
Production APIs with real concurrent traffic. My work sat on the request path: how services behave when load spikes, how data gets fetched without choking the database, how failures get caught before they cascade. Started integrating LLM backed features into live services too: wrapping model calls behind APIs, handling streaming responses, and making sure AI latency did not tank the rest of the stack.
Jun 2022 to Dec 2023
03Software Engineer
Living inside the event stream
Most of my time went into Kafka pipelines and the systems around them. Tracing events through production, figuring out why a consumer group lagged at 2 AM, tightening retry paths so bad data never poisoned downstream services. It was backend engineering at enterprise scale: less about shipping features fast, more about making sure what already existed did not break when the world got loud.
Projects
Personal systems work. Same engineering instincts, smaller scope.
01
Distributed transaction handling
Built around idempotent requests, retry logic, and keeping payment state consistent when any step in the chain fails. Focused on the backend orchestration layer, not the UI.
02
Shared traffic control
Centralized rate limiter using Redis sliding windows. One place to define policies, one place to decide allow or deny before bad traffic hits core services.
03
Background work that finishes
Distributed scheduler with retry backoff, dead letter handling, and visibility into stuck jobs. Built for the case where a worker dies mid task and nothing gets lost.
04
AI inference in production
Backend layer for LLM calls in product features. Streaming responses to the client, token budgeting, response caching, and fallbacks when models slow down or rate limit. AI treated like any other service dependency.
Tools & Skills
Grouped by area. The resume has the full list tailored to each role.
Languages
Backend
Frontend
Cloud & DevOps
Data & Storage
AI & ML
Education
Graduate
Master of Science, Computer Science
Where I went deep on algorithms, concurrency, systems design, and machine learning fundamentals. The theory behind the production work, including the AI systems I build today.
Undergraduate
Bachelor of Technology, Computer Science
Foundation in computer science. Data structures, programming, and the first principles everything else builds on.