77-day longest streak
Tirth Kanani I am an AI engineer and technical founder based in London. I build production-grade tools and products that developers actually use. I created code-review-graph, an open-source developer tool…
Tirth Kanani
I am an AI engineer and technical founder based in London. I build production-grade tools and products that developers actually use. I created code-review-graph, an open-source developer tool that has reached over 12,000 GitHub stars and 1,400 forks. Shipped as a Claude Code plugin with full MCP server integration, it uses Tree-sitter AST parsing across 23 languages to build a persistent incremental knowledge graph. It re-indexes a 2,900-file monorepo in under two seconds and delivers 6.8 times average context compression, up to 49 times on large monorepos. I published reproducible benchmarks on real repositories including httpx, FastAPI and Next.js, added a D3.js visualisation, and maintain an active community that shapes its roadmap.
As sole technical founder I also shipped CrumbleUX, a real-time VLM product for design critique, from initial idea to paying customers in weeks. I worked in the University of Birmingham HCI and AI Lab building evaluation pipelines and Graph Neural Networks for multi-agent reliability.
I won first place at the Epiminds Multi-Agent Hackathon 2025 with SupplyMinds and open-sourced MERIT and Jailbreak-Eval, my evaluation and red-teaming frameworks.
Get in touch at [email protected]. You can also find me on LinkedIn, on my website, or on GitHub.
Building reliable AI that scales — always iterating.
-
GraphMinds ★ PINNED
GraphMinds: A Python project exploring graph-based algorithms and AI models.
Python ★ 8 1y agoExplain → -
code-review-graph ★ PINNED
Local-first code intelligence graph for MCP and CLI. Builds a persistent map of your codebase so AI coding tools read only what matters, with benchmarked context reductions on reviews and large-repo workflows.
Python ★ 19k 6d agoExplain → -
claude-games ★ PINNED
No description.
TypeScript ★ 3 2mo agoExplain → -
ResearchWebGraph
AI tool to explore papers, build knowledge graphs, and get LLM answers with vector search. Built with FastAPI and Streamlit. Open-source.
Python ★ 7 1y agoExplain → -
tirth8205
My Personal Repository
★ 5 1mo agoExplain → -
Jailbreak-Eval
Production-grade LLM red-teaming framework with multi-agent swarm, GCG, PAIR, mutation engine, and Streamlit dashboard — for defensive AI safety research only.
Python ★ 5 2mo agoExplain → -
EEGSpeech
A brain-computer interface (BCI) for decoding speech phonemes from EEG signals using a hybrid CNN-LSTM model, with interactive Streamlit visualizations and Docker support.
Python ★ 3 11mo agoExplain → -
Understand-Anything ⑂
Graphs that teach > graphs that impress. Turn any code into an interactive knowledge graph you can explore, search, and ask questions about. Works with Claude Code, Codex, Cursor, Copilot, Gemini CLI, and more.
TypeScript ★ 2 1d agoExplain → -
RAG_using_NLP
A local GPU-accelerated Retrieval-Augmented Generation (RAG) pipeline for PDF question-answering with multi-LLM support and modular NLP components. Process documents locally with privacy-focused information retrieval.
Python ★ 2 1y agoExplain → -
my-portfolio
A modern, responsive portfolio website showcasing my skills, projects, and experience as a Software and Machine Learning Engineer.
TypeScript ★ 1 1mo agoExplain → -
CV2Latex
No description.
TypeScript ★ 1 5mo agoExplain → -
MERIT
No description.
Python ★ 0 3mo agoExplain → -
transformers ⑂
🤗 Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.
★ 0 4mo agoExplain → -
SafetyVision
No description.
Python ★ 0 11mo agoExplain → -
Landmarks
No description.
Swift ★ 0 5y agoExplain → -
MoodMoji
MoodMoji is a Python-based desktop application that uses AI to detect emotions from webcam footage in real-time, mapping them to emojis for a fun, interactive experience. It features emotion recognition, custom emoji capture, and mood-based Spotify playlist suggestions, built with OpenCV, TensorFlow, and Tkinter.
Python ★ 0 1y agoExplain → -
Accimap-Analysis
No description.
Jupyter Notebook ★ 0 1y agoExplain → -
Project_Mobile-and-Ubiquitous-Computing
Integrated System for Smart Street Light Management, Urban Traffic Mapping, and Car Parking Booking Management
★ 0 1y agoExplain → -
Walkwise
No description.
★ 0 1y agoExplain → -
Airline-management-system
No description.
HTML ★ 0 5y agoExplain → -
CSE518_ArtificialIntelligence
No description.
Jupyter Notebook ★ 0 1y agoExplain → -
Rotation-of-Cube-and-Sphere-using-Concepts-of-LInear-Algebra
No description.
Java ★ 0 5y agoExplain → -
MarkovDecisionProcess
No description.
MATLAB ★ 0 5y agoExplain → -
Audio-Encryption
No description.
MATLAB ★ 0 5y agoExplain → -
Weather-Data-Analysis
No description.
MATLAB ★ 0 5y agoExplain → -
Smart-Street-Light-System
Smart Street Light Management System Using Arduino Uno and IR Sensors
C++ ★ 0 5y agoExplain → -
Computer-Networks-Project_Internet-Radio
No description.
C ★ 0 3y agoExplain → -
Volatility-Prediction_Machine-Learning-Project
No description.
Jupyter Notebook ★ 0 3y agoExplain → -
Alumni-reinvited
No description.
HTML ★ 0 5y agoExplain → -
mathematics-for-machine-learning ⑂
Mathematics for Machine Learning specialization from Coursera platform
★ 0 6y agoExplain →
No repos match these filters.