Machine Learning
Training and evaluating deep learning models (CNNs, Transformers, XGBoost) with SHAP explainability.
Click to learn moreWelcome to my portfolio
Hi, I'm Shiv
AI/ML Engineer
I build intelligent systems — from deep learning models and computer vision pipelines to LLM-powered applications. Let's build something smart together.
About Me
Hi, I'm Shiv! A Computer Science undergrad minoring in AI who builds and ships Python-based AI systems end-to-end — from integrating LLM APIs into production apps to training deep learning models evaluated with rigorous, production-relevant metrics. When I'm not training models, I'm exploring RAG pipelines, agentic workflows, and open-source AI tooling.
LLM Engineering
Building with LLM APIs, prompt engineering, RAG, and agentic workflows (LangChain/LangGraph).
Click to learn moreFull-Stack for AI
Shipping AI features with FastAPI, PostgreSQL/Redis, and React/Streamlit front ends.
Click to learn moreAlways learning, always building 🚀
Skills & Tech Stack
The tools I use to research, build, and ship reliable AI products.
LLM & RAG
Machine Learning
Production AI
GitHub Activity
My Journey
June 2024 – June 2028
B.Tech, Computer Science & Engineering (Minor in AI)
Pranveer Singh Institute of Technology (PSIT), Kanpur
Started my CS degree with a minor in Artificial Intelligence.
Patent Published
Patent Published, Government of India
ML-Based Aquaculture Health Prediction & Decision Support System
Engineered domain-specific features and trained Gradient Boosting/XGBoost models (R² 0.91), with an interactive Streamlit app and SHAP explainability.
May 2026 – Present
Research Intern → Undergraduate Student Researcher
Indian Institute of Technology (Banaras Hindu University), Varanasi
Designed ACS-SegNet, a hybrid CNN–Transformer segmentation architecture, achieving 92.04% accuracy; deployed as a FastAPI inference service.
Next
Looking for my next opportunity!
Ready to contribute to a team building useful, reliable AI systems.
Projects
From AI systems to dependable product infrastructure, I enjoy solving practical problems with code.
AlphaPlus — AI-Powered Crypto Futures Trading Bot
AI-driven trading system using 14+ technical indicators, vectorized backtesting, risk metrics, and Binance API signal execution.
ACS-SegNet — Histopathology Segmentation
Hybrid CNN–Transformer tissue-segmentation architecture achieving 92.04% accuracy, deployed as a production FastAPI inference service.
AutoWebsite-Maker
AI-powered landing-page personalizer using Gemini API for ad-intent extraction, Playwright + BeautifulSoup, FastAPI, React, and Docker.
AnyDoc — AI Document Analyzer
LLM-powered document summarization, key-insight extraction, and Q&A with PDF parsing and chunked processing.
SniURL — Full-Stack URL Shortener
FastAPI + React application with JWT auth, PostgreSQL/SQLAlchemy, async Redis caching, pytest, and Docker Compose.
Aquaculture Health Prediction
Patent-published Government of India system using Gradient Boosting/XGBoost models (R² 0.91) and SHAP explainability.
DetailsPatent Published
Impressed? Let's have a talk!
Building AI systems and always open to interesting problems — let's talk.
Let's Talk