Designed and implemented an LLM-powered chatbot to schedule appointments, capable of understanding and classifying 50+ distinct user sentiments for more natural and context-aware conversations.

Built end-to-end conversational pipelines in Python, integrating NLP models, embeddings, and sentiment classifiers with downstream scheduling APIs.

Applied prompt engineering and memory/state management techniques to maintain multi-turn conversations.

Designed and deployed a collaborative filtering recommendation system on AWS (Lambda + API Gateway), driving 12% conversion uplift and 15% longer session durations.

Built feature pipelines on Apache Spark (EMR) using user interaction, product catalog, and review streams from Kafka + S3, generating ML-ready behavioral and product embeddings.

Implemented matrix factorization models in PyTorch (ALS) with automated retraining workflows, optimized for low latency and top-k recall accuracy.

Developed Streamlit dashboards to monitor CTR, recall@k, and NDCG, enabling cohort-based evaluation and actionable insights for stakeholders.

Integrated model outputs into production via FastAPI microservices (Docker, Kubernetes, Redis cache) and collaborated on A/B testing frameworks to validate performance

Re-architected legacy PHP/MySQL monolith into MERN-based microservices, enabling horizontal scalability for 50k+ concurrent users and reducing system bottlenecks.

Designed and deployed RESTful APIs with Node.js + Express, integrating real-time shipment tracking and transactional services with low-latency SLAs.

Implemented CI/CD pipelines (GitHub Actions) with automated unit/integration tests, achieving faster, safer deployments and reducing release errors.

Optimized database design and query performance, improving response times by 30% and ensuring reliability during peak traffic.

Migrated a server-rendered app to a React-based SPA, improving load times and maintainability.

Contributed TypeScript-based components and assisted in building modular frontend dashboards.

Participated in Agile sprints delivering scalable features aligned with engineering best practices.