Third-year CS student (B.E. 2028) β’ AI Software Engineer β’ Building production-shaped AI systems
Software Engineering Intern / AI Software Engineer Intern β targeting roles that convert to full-time SWE / GenAI engineering positions.
Open to: Remote β’ Relocation (Bengaluru, Hyderabad, Pune) β’ On-site
| Superpower | Evidence |
|---|---|
| End-to-end AI system building | RAG pipelines, multi-agent orchestration, evaluation frameworks |
| Fast AI prototyping | Idea β working system in days, not weeks |
| Critical judgement on AI output | Code review, design docs, validation before merge |
| Production-shaped thinking | Shipped tested PRs in live client codebase at Indium Software |
π€ AgencyOS β Self-healing agentic AI platform
Multi-agent, event-driven platform of 5 LangGraph-orchestrated microservices coordinated via Agent-to-Agent (A2A) protocol
Self-healing web-intelligence layer with deterministic health scoring + automated schema-drift diagnosis/repair
Python LangGraph A2A Protocol FastAPI PostgreSQL Docker
π RAG-Based Intelligent Document QA β Production RAG pipeline
Semantic chunking β’ Async document ingestion β’ Scalable embedding generation
Hybrid retrieval: BM25 + dense vector + MiniLM cross-encoder reranking
Grounded, context-aware answers with evaluation harness
Python LlamaIndex ChromaDB FastAPI MiniLM BM25
π§© CanonPulse β Agentic narrative intelligence on Databricks
Dual-layer temporal graph reasoning over serialized fiction
Evidence-first pipeline: LLM extraction β deterministic ledger resolution
269-test offline-first validation suite
Python Databricks Graph DB LangGraph Testing
ποΈ DepthWizard β Single-view 3D reconstruction
Depth Anything V3 backbone + physical solar-shadow verification
Calibrated uncertainty via split conformal prediction
GPU clipmap rendering for real-time terrain visualization
Python PyTorch CUDA Computer Vision Conformal Prediction
AI YouTube Shorts Generator (4.2k β / 750+ π΄)
PR #57 merged β Fixed transcription pipeline: 0 β 59 segments output
Contributed to a viral open-source AI content creation tool
Python Whisper FFmpeg Video Processing
AI/ML Intern @ Indium Software (Nov 2025 β Jan 2026)
- Built RAG pipeline for Liftr (AI legacy code modernization platform) β ingested source code, README, TOML, YAML, configs at scale
- Shipped production-quality PRs in live client-facing codebase with senior engineer code review
- Authored technical design docs for architecture decisions; validated AI-generated output critically
Languages: Python Java C++ C JavaScript Go TypeScript SQL
AI/ML: PyTorch Hugging Face Scikit-learn LLMs RAG LangChain LlamaIndex Vector Embeddings
Agentic AI: LangGraph Multi-Agent Systems A2A Protocol Agent Orchestration Tool Use Prompt Engineering
Backend: FastAPI Django Node.js Socket.IO PostgreSQL MongoDB ChromaDB
Platforms: Docker Docker Compose Playwright Linux AWS Azure GCP GitHub Actions
Note: GitHub stats widgets may occasionally show placeholder images when the external services are temporarily unavailable. The contribution graph above is served from a separate service.
- π₯ NextGenHacks Winner β Best Idea Track
- π 650+ LeetCode problems solved (DSA rigor)
- π Technical writing β Design docs, architecture decisions, evaluation reports
- π¬ Research mindset β Offline-first validation, conformal prediction, deterministic verification
"Builder who pairs learned models with deterministic, verifiable reasoning" β the failure mode that drives my work: systems that look healthy while producing nothing.

