I’m Esakkiraja
Full-Stack AI Engineer
What I actually do
Full-stack development
End-to-end applications — React and Next.js frontends on FastAPI and Express backends, with REST APIs, PostgreSQL and MongoDB schemas designed, shipped and maintained.
LLM infrastructure
A shared multi-provider framework across OpenAI, Claude, Gemini, Gemma, Llama and Mistral — model abstraction, token accounting and parameter control, plus RAG pipelines over vector databases.
Guardrails
Multi-layer safety guardrails for AI assistants, backed by a telemetry and evaluation pipeline measuring per-guardrail precision and recall in production.
Docker & Kubernetes
Containerised services deployed and operated on AWS and GCP — Docker images, Kubernetes workloads and CI/CD pipelines from build through post-release monitoring.
Testing & quality
Unit and integration tests around services and API contracts, run in CI so regressions surface before release — alongside evaluation suites that keep AI behaviour measurable.
AI-assisted development
Building and integrating with agentic coding tools — Claude Code, GitHub Copilot and OpenAI Codex — to accelerate delivery across the stack.
Where I’ve shipped
Two and a half years at Skill Rank building AI systems in production, starting from an internship.
Software Engineer
Jul 2024 — Present- Built production AI microservices and a shared multi-provider LLM framework supporting OpenAI, Claude, Gemini, Gemma, Llama and Mistral — with model abstraction, token accounting and parameter management, deployed on AWS (ECS, ECR, Lambda, S3, CloudFront, Route 53, IAM).
- Architected RAG pipelines using LangChain, LlamaIndex, CrewAI, embeddings, Jinja2 templates and vector databases (Milvus, Chroma, pgvector, Pinecone); implemented a multi-threaded Python code generation agent.
- Designed and deployed 45 AI guardrails for a patient-facing healthcare assistant serving ~240K requests/month, backed by an evaluation pipeline over 1.65M+ decision events for precision, recall and threshold tuning.
- Enabled multimodal AI with Gemma 4B, Hugging Face, fine-tuned open-source LLMs, OCR, vision models and Whisper — serving 1M+ requests while cutting document processing latency.
- Designed Neo4j knowledge graphs with millions of nodes, and used Claude Code, Cursor and OpenAI Codex to accelerate AI-assisted development.
Software Engineer — Intern
Jul 2023 — 2024- Orchestrated async FastAPI RAG pipelines using Azure OpenAI, OpenAI, Claude, vector databases, embeddings, function calling and structured JSON outputs — cutting enterprise chat query latency by over 50% (6s to under 3s).
- Developed enterprise AI chat agents on Microsoft Bot Framework with semantic search, handling 20+ concurrent users, using Azure Blob Storage and Azure Key Vault for secure document and secret management.
- Built an AWS document intelligence pipeline with SQS, Textract and MongoDB automating OCR invoice extraction with human validation — 100+ invoices/day at 95% extraction accuracy.
Full Stack Developer — Intern
Feb 2022 — 2023- Created an end-to-end full-stack platform — responsive Bootstrap frontend, backend services, REST APIs and optimised MySQL schemas — supporting 1,000 concurrent users.
- Resolved 40+ production bugs, took part in Agile sprint planning and delivery, and integrated third-party APIs to improve stability and development velocity.
The toolkit
Languages
Frameworks
AI / LLM
Data
Cloud & DevOps
Engineering
Things I built and shipped
CreateShorts
createshorts.in ↗Text prompt → finished short-form video
An AI video generation pipeline that turns a single text prompt into a finished short — orchestrating image generation, text-to-speech voiceover, animation, sound effects and automated rendering into one flow.
TalkViz
talkviz.in ↗Plain English → SQL and charts
A natural-language data analysis platform converting plain-English questions into SQL and interactive visualisations. 6 data sources across live PostgreSQL/MySQL/MongoDB connections and CSV/Excel/JSON uploads, monetised across 3 pricing tiers with usage-based query limits.
Where it started
B.E. Computer Science & Engineering
Francis Xavier Engineering College, Tirunelveli
🏅 Buddy Professional Award, 2023
Let’s build something worth shipping
Open to conversations about AI engineering, LLM infrastructure and product work. Based in Tirunelveli, Tamil Nadu — working remote.