Final-year CSE student at IIIT Kottayam. I like backend systems, and lately I've been especially interested in the parts of software that are hardest to make reliable. That's how I ended up working on agentic AI during my internship.
Internship @ Twinleaves. I worked on the Task Management feature of an AI-powered retail chatbot (Java, Spring, Google ADK, BigQuery). I designed the conversation flow the LLM follows to create and track tasks, wired it into three backend services, and (the part I'm genuinely proud of) replaced a flaky setup where the LLM wrote its own database joins with a deterministic resolution layer in Java, so the output is correct every time instead of most of the time. I also tracked down a subtle bug where the model was mishandling its own tool output, by pulling the raw framework payload and tracing it back to the actual cause. Before this, I built a Google Merchant Center onboarding platform (Java, Spring Boot, PostgreSQL, OAuth 2.0) that turned a fully manual process into an automated one.
Adavya Retroverse, Live Point Tracker. A real-time leaderboard for a college carnival, with two-tier role-based admin access and live score updates. Built and shipped under a tight deadline, and actually used by people during the event, which is a different kind of pressure than a class project. Went back after the event to fix an auth vulnerability I found, replacing it with proper Supabase Auth and RLS policies. Stack: React, Supabase, Tailwind, Vercel Repo
Carbon Emission Tracking System. A full-stack app I built solo to track and analyze carbon emissions across regions and sectors, with role-based auth, an admin dashboard, and visualizations to make the data actually readable. Stack: Flask, MySQL, Matplotlib, Seaborn Repo
AWS Guided Labs. Notes and configs from guided mini-projects across EC2, S3, Lambda, VPC, IAM, RDS, API Gateway, DynamoDB, and QuickSight. Not production experience, just me working through the services hands-on and writing down what broke and what I learned. Repo
- Deep in placement prep: DSA, competitive programming (300+ LeetCode, Codeforces 862), and system design
- Building a full-stack project with Node.js and React, end to end, including containerization, AI features, and AWS deployment
- Open to backend, full-stack, or applied-AI internship opportunities
Core: Java, Python, JavaScript, SQL, Spring Boot, Flask, REST APIs, PostgreSQL, MySQL, Supabase, Git
Exploring: AWS, Docker, Kubernetes