Skip to content
View Sehaan-1's full-sized avatar
3 followers · 4 following

Achievements

Achievement: QuickdrawAchievement: Pair Extraordinaire

Achievements

Achievement: QuickdrawAchievement: Pair Extraordinaire

Block or report Sehaan-1

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Sehaan-1/README.md

👨‍💻 About

CS student at VIT Bhopal building applied AI systems that are reproducible, testable, and measurable.

What I optimize for:
✓ Evaluation-first ML (baselines, metrics, ablations)
✓ Reliability-minded engineering (idempotency, recovery, clean interfaces)
✓ Practical deployment (Docker, CI, sensible defaults)

Roles I'm targeting:
AI/ML Engineer • Data Scientist • MLOps-minded roles

🚀 Featured Projects

🏛️ Tender Royal Pulse

Resilient public procurement crawler

Extracts 50,000+ tenders from India's eProcure portal using Playwright with session persistence, heartbeat monitoring, and idempotent upserts.

Tech: Python • Playwright • Pydantic
Highlights: Retry taxonomy, graceful shutdown, strict typing

⚖️ Cross-Border Legal RAG

Hybrid retrieval + rule layer + evaluation

Answers legal questions using EU regulations via BM25 + FAISS hybrid retrieval, reranking, and refusal guards.

Tech: Python • FAISS • BM25
Metrics: 78% article recall, 85% refusal accuracy, ~9s p95 latency

🎥 SAR — Suspicious Activity Recognition

Real-time CV surveillance platform

Multi-camera security system with React dashboard, Node.js backend, and Python AI worker (YOLOv8 + ByteTrack).

Tech: React • Node.js • OpenCV • Docker Compose
Highlights: WebSocket alerts, dynamic threading, one-command deploy

🛡️ XAI Fraud Detection

FastAPI model serving + SHAP explainability

Fraud detection with XGBoost + SMOTE, serving via FastAPI with per-prediction SHAP attributions and Prometheus-style monitoring.

Tech: Python • XGBoost • SHAP • FastAPI • Docker
Highlights: CI/CD pipeline, /metrics endpoint, cloud-ready

🛠️ Stack

Core
Python TypeScript JavaScript Java R



AI / CV / Data
OpenCV SQLite



Web
React Node.js HTML5 CSS3



DevOps / Cloud
Docker Git AWS GCP

📊 GitHub Stats

streak
Footer

Pinned Loading

  1. Tender-Royal-Pulse Tender-Royal-Pulse Public

    A full-stack dashboard and crawler for government tenders. Built with FastAPI and React to make finding and tracking public procurement data easier and more accessible.

    Python

  2. crossborder-legal-rag crossborder-legal-rag Public

    A Retrieval-Augmented Generation (RAG) system for EU Private International Law. Combines hybrid retrieval (BM25 + FAISS) and a deterministic rule engine to answer cross-border legal queries with of…

    Python

  3. sar---suspicious-activity-recognition sar---suspicious-activity-recognition Public

    AI-powered Suspicious Activity Recognition system — React + TypeScript frontend, Node.js/Express backend, Python YOLOv8/CLIP AI worker, Docker Compose

    TypeScript

  4. xai-fraud-detection xai-fraud-detection Public

    An Explainable AI (XAI) fraud detection engine built as an MLOps project. Combines an XGBoost model with SHAP explanations, served via a FastAPI backend and an interactive analyst console.

    Jupyter Notebook

  5. docker-incident-controller docker-incident-controller Public

    An autonomous SRE agent for Docker. Watches Docker containers for known failure signatures, creates SQLite-tracked incidents, and executes a multi-step remediation plan. API + Prometheus for observ…

    Python