Hari Sai Kaja

Architected an AI-driven clinical simulation platform end to end on a Next.js and Supabase (PostgreSQL) stack and piloted it with nursing faculty. Built an agentic AI layer using MCP and LangGraph over Claude and Ollama, with a RAG pipeline grounded in prior cases via pgvector (HNSW). Delivered a faculty review workflow, implemented CI/CD with automated tests, and deployed continuously to AWS and Kubernetes.

Engineered a lightweight distributed consensus protocol (PoEWAL) for resource-constrained IoHT edge systems and implemented the consensus logic in Go. Built event-driven pipelines across Layer 2 and Layer 3 using Kafka and pub/sub messaging with secure authentication. Delivered an end-to-end Azure cloud-native data pipeline using Data Factory, Blob Storage, and RBAC-secured serverless functions, and created benchmarking automation in Python and Shell.

Modernized an aging PHP/React monolith into a Next.js and NestJS TypeScript architecture and served multiple enterprise clients. Improved performance by migrating the data layer from MySQL to PostgreSQL with Prisma ORM, schema modeling, and query/index optimizations. Built consistent UI with React/Next.js and Tailwind/shadcn, shipped key payroll functionality end to end, and improved delivery using CI testing and AI-assisted development. Managed AWS infrastructure with Terraform and deployed containerized services on EKS while designing OAuth2/JWT/HMAC-based multi-tenant access control.

Ran distributed event-driven Kafka pipelines to power election and COVID-19 dashboards using ETL and KDD workflows. Built FastAPI/Flask REST APIs and React/Next.js dashboards deployed on Vercel for newsroom usage under tight deadlines. Designed a low-latency analytics platform on Apache Druid, tuned queries for sub-second responses, and implemented scraping and ingestion with BeautifulSoup and Scrapy backed by Redis caching and MongoDB/Elasticsearch/Firebase storage. Maintained observability with New Relic and administered Hetzner bare-metal infrastructure with Nginx and Ansible.

Built backend services and REST APIs in Go alongside a Python and Node.js stack for customer-facing products. Shipped the TaxFriday AI chatbot with a React frontend, Python/Flask and Node.js backend, and owned platform DevOps on AWS including EC2 IAM/CloudWatch and load balancing, plus Docker and Ansible automation and Jenkins-based delivery. Implemented a high-throughput NLP sentiment pipeline using Elasticsearch, Kafka, Redis, and MongoDB, including query and indexing optimizations and caching for read-heavy paths. Embedded DevSecOps scanning into CI and used Prometheus/Grafana for observability with on-call operations.

Built a full-stack Django/Flask inventory management system using MySQL and GitLab workflows, including an Android client for real-time order tracking and automated reporting. Developed an IoT pollution-monitoring dashboard using NodeMCU telemetry and implemented edge ingestion in Embedded C over MQTT. Supported the delivery and maintenance of features by converting stakeholder workflows into working functionality within a small production team.