Final-year B.Tech Computer Science student
Interested in AI/ML, generative AI, and software engineering. I learn by building projects and exploring research problems.
- Machine learning and deep learning
- Generative AI and agent-based applications
- Applied research and reproducible experiments
- Backend development and practical software engineering
| Project | What it is |
|---|---|
| CSD - Clustering Stability-Accuracy Discordance | Co-authored research software and evidence workflow studying why resampling stability can diverge from agreement with external labels in high-dimensional biomedical data. |
| ICC Clustering |
Python research software for consensus clustering, anchor-graph methods, dataset handling, and comparative evaluation. Indexed in the XERJ Corpus Hub as a reference corpus for AI agent navigation and AST-aware search. See its evidence status. |
| NullConsensus | Follow-on research implementation exploring permutation-null feature calibration and count-aware gating, with a frozen pre-registration protocol. |
| OphthalmoAI | Retinal fundus screening research prototype combining deep vision models, probability calibration, Grad-CAM explainability, and browser-side ONNX inference. Not a clinical diagnostic device. |
| Project | What it is |
|---|---|
| C.O.P.P.E.R. | Local-first personal AI runtime exploring specialized agents, routing, persistent memory, safety controls, and a desktop interface. Research prototype; see the benchmark-validity review before interpreting its reported results. |
| AI-SATHI | Voice-first financial assistant for Self-Help Group workflows, with Bengali voice processing, structured ledger extraction, grounded assistance, and report generation. Developed as part of an applied AI/social-innovation project. |
| Agentic Research Framework | Reproducible research workflow and agent-governance starter kit for multi-agent coding tools, experiment tracking, evidence-linked claims, and manuscript synchronization. |
- SpecBridge - hardware/app compatibility and recommendation platform.
- ResumePilot - AI-assisted resume workflow.
- DemoTradingBot - Python CLI for Binance Futures demo trading, with Decimal-based order validation and dry-run support.
- TenantAPI - tenant-scoped API and data-access example.
- Languages: Python, JavaScript, TypeScript, Java, SQL
- AI/ML: PyTorch, scikit-learn, LangGraph
- Development: FastAPI, React, PostgreSQL, Docker, Git
- Microsoft Certified: Machine Learning Operations Engineer Associate (AI-300)
- Microsoft Certified: Azure AI Apps and Agents Developer Associate (AI-103)
- Oracle Certified: Agentic AI Foundations Associate
- Oracle Certified: AI Database Foundations Associate
- Treat benchmarks as hypotheses to verify, not marketing numbers.
- Keep research claims linked to code, data provenance, and reproducible evidence.
- Design for failure modes, security boundaries, and human review-not only the happy path.
- Prefer clear limitations over unsupported claims.
Currently: Final-year B.Tech CSE student, building projects and preparing for opportunities in AI/ML and software development.


