• Owns end-to-end volunteer operations for the Bangalore chapter — recruitment, onboarding, coordination, and retention
• Responsible for reliably deploying volunteers across food redistribution drives and building the systems that make a distributed team show up and deliver

• Designed and shipped a full-stack payment platform — REST API, real-time webhook delivery with retry logic, recurring subscriptions, multi-currency support, and a live analytics dashboard
• Handled horizontal scalability with Kubernetes and end-to-end observability via Prometheus/Grafana

→ 1,500+ TPS peak · 45 ms avg response · 99.9% uptime · 95%+ test coverage across 50+ cases

• Built a serverless AWS pipeline that ingests bulk applicant data via S3, scrapes available loan products, runs AI-assisted eligibility matching, and delivers personalised notifications — no human in the loop
• Decoupled into three independent n8n workflows (discovery → matching → notification) for modularity and reliability

→ Fully automated applicant-to-offer pipeline: CSV upload to personalised email, zero manual steps

• Top competitors clustered at 96–97%; built a systematic multi-model ensemble (XGBoost, AutoGluon, CatBoost, SVMs, neural nets) and analysed model disagreement to find the performance edge

→ 97.5% accuracy — top of leaderboard in a dense competitive field

• Implemented Dueling Double DQN with Prioritised Experience Replay — separating value/advantage estimation and focusing training on high-TD-error transitions to stabilise learning from raw pixels

→ Agent learns a stable jump policy from visual input alone; ships with pre-trained weights