- 🎓 B.Tech CSE at IIIT Senapati, Manipur (2023–2027)
- 🔧 Currently building NovaVCS — a version control system from scratch in C++17, with a custom DAG-based commit graph, Myers diff engine, and Trie/inverted-index search
- 🧠 Recently shipped OmniScan AI — a fully local vulnerability scanner combining SAST/DAST engines with a RAG pipeline (FAISS + local LLM via Ollama) for context-aware remediation
- 🌾 Also built AGRIHIVE — an AI farming advisory platform with a self-trained CNN (92% accuracy) and Gemini-powered multilingual chatbot
- 📊 Interned as a Data Analyst at Varun Beverages Limited, building churn prediction pipelines (XGBoost, Random Forest) and market basket analysis (Apriori, FP-Growth)
- 🏆 Global Contest Rank 1623 on LeetCode
- 📫 Reach me at harshithihu@gmail.com
NovaVCS — High-performance version control system in C++17 Content-addressable storage, three-way merge with JSON-serialized conflict states, Huffman + delta compression for packfiles, sub-millisecond fuzzy search via Levenshtein automata.
OmniScan AI — Autonomous vulnerability scanner (Python, Node.js, React) 100% local and privacy-centric. RAG pipeline over OWASP knowledge base, attacker-persona simulation, and LLM-generated remediation code for chained vulnerabilities.
AGRIHIVE — AI farming advisory platform (Flutter, Firebase, TensorFlow) Offline-first architecture serving 100+ concurrent rural users, plant disease detection via custom CNN, multilingual chatbot via Gemini API.
Languages: Python, C++, JavaScript, SQL
Backend: Node.js, Flask, REST APIs, Firebase, Docker
Data & ML: Pandas, NumPy, XGBoost, scikit-learn, TensorFlow, CNNs, NLP, LLMs, RAG pipelines, Prompt Engineering
Tools: Git, Linux, Postman, Jupyter, Streamlit
