• Designed and integrated AI agents and RAG pipelines using Pinecone and ChromaDB with GPT-4 and Claude API - enabling context-aware automated query handling that reduced manual document triage time by 55% across enterprise operations workflows.

• Architected event-driven microservices using async message patterns and loosely coupled service boundaries - eliminating cascading failures during peak load and enabling independent scaling across distributed cloud-native services.

• Provisioned and managed AWS cloud infrastructure (EC2, S3, Lambda, RDS) including serverless event-driven jobs and auto-scaling policies - maintaining 99.5%+ pipeline availability and cutting infrastructure provisioning time by 40% through IaC automation.

• Implemented backend microservices in Node.js and Python for LLM API orchestration with circuit breakers, retry logic, adaptive rate limiting, and request caching - achieving sub-400ms p95 response times under variable LLM latency conditions.

• Designed relational and vector database schemas for high-throughput operations; built document preprocessing pipelines for chunking and embedding generation - improving semantic search retrieval precision by structuring context for downstream LLM generation.

• Built real-time and batch cloud data pipelines connecting event streams to downstream cloud storage and analytics layers - ensuring data consistency and reducing pipeline lag to under 2 seconds for time-sensitive operational data.

• Containerized services with Docker and maintained Prometheus/Grafana/Kibana observability stack with structured alerting - reducing mean time to detect(MTTD) incidents by 50% compared to prior reactive monitoring approach.

• Maintained and monitored university department websites, ensuring 99%+ uptime

• Implemented WCAG AAA accessibility compliance using Stark and AChecker tools

• Assisted faculty with building research project websites and testing environments

• Improved website accessibility for users with disabilities (color contrast, alt text, keyboard navigation)

• Mentored 150+ students in Data Science and Big Data courses

• Taught Python, SQL, Pandas, and data analysis techniques

• Facilitated hands-on labs on distributed data processing (Spark, Kafka)

• Conducted code reviews and provided feedback on programming assignments

• Diagnosed and resolved a critical production WebSocket outage (85% error rate, 200+ users affected): traced reconnection storms via error log and network traffic analysis; implemented connection pooling, retry logic, and heartbeat mechanism - restoring stability within hours and dropping error rate below 5%.

• Reduced API response time by 75% (800ms to 200ms) on high-traffic endpoints: identified missing indexes on frequently-joined MySQL tables, refactored query patterns, and tuned connection pool settings - fully resolving customer-facing performance complaints and recovering SLA compliance.

• Built monitoring dashboards tracking API health, latency, error rates, and throughput; participated in on-call rotation with structured post-mortems - achieving zero repeat incidents from previously resolved failure modes across the production lifecycle.

• Enforced SQL injection prevention, JWT authentication, and input validation throughout the codebase; wrote unit and integration tests (JUnit, Jest) achieving high coverage across all critical business logic and API endpoints.

• Collaborated with product, frontend, QA, and infrastructure teams defining API contracts, translating requirements into deliverables, and communicating incident status clearly to non-technical stakeholders.