I am moving from senior data analytics into applied AI engineering. I build testable systems around retrieval, agent workflows, analytics, and developer automation, with an emphasis on evidence, evaluation, and guardrails.
| Project | Engineering proof | Stack |
|---|---|---|
| citipoints | Retail loyalty analytics with ML models and an auditable natural-language-to-SQL layer over synthetic data | TypeScript, Next.js, FastAPI, XGBoost |
| rag-analytics-assistant | Guardrailed analytics assistant with hybrid retrieval and tool-calling over user-controlled data | Python, FastAPI, DuckDB, RAG |
| codex-upgrade | Portable Codex CLI skills, prompts, and operating guardrails | Python, Codex CLI |
| claude-code-toolkit | Hooks and reusable controls for session management and credential protection | JavaScript, Claude Code |
| officehub | Self-hosted project workspace with timelines and management rollups | TypeScript, Next.js, Postgres, Docker |
| ai-engineering-receipts | Public interview playbook and evidence index for applied-AI work | Documentation, evidence maps |
Six merged code pull requests, verified against the upstream repositories:
| Repository | Contribution |
|---|---|
| gradio-app/gradio#13396 | Convert non-WAV audio output to int16 to prevent playback corruption |
| gradio-app/gradio#13344 | Fix a KeyError involving nested gr.Examples inside gr.render |
| livekit/agents#5288 | Make function-call history preservation configurable in AgentTask |
| mem0ai/mem0#4461 | Add reasoning_effort support for reasoning models |
| optuna/optuna-integration#278 | Fix late binding in BoTorch constraint closures |
| Mintplex-Labs/anything-llm#5560 | Strip Markdown before text-to-speech processing |
Applied AI : RAG, agent workflows, tool calling, evaluation, guardrails
Backend : Python, FastAPI, Node.js, PostgreSQL
Frontend : TypeScript, React, Next.js
Data : Pandas, SQL, forecasting, segmentation, experimentation
Delivery : Docker, CI, reproducible validation, evidence-first documentation
- Building public, reviewable AI-engineering proof rather than tutorial clones.
- Contributing tested fixes to the open-source tools used in real projects.
- Looking for applied AI, AI engineering, and AI analytics roles open to remote work from India.


