Fawaz AI safety, efficient AI, and intelligent software

Featured Work

Trace quietly remembers what you do on your computer: files touched, tabs opened, commands run, and lets you ask about any of it in plain language. 100% local, private memory engine and context layer for AI assistants. trace.zenyslab.com

What I Do
AI Safety
Working on making AI systems more reliable, interpretable, and aligned with human intent, with a focus on understanding model behavior and building safeguards into real-world deployments.
Efficient AI
Reducing the compute, memory, and energy cost of AI workloads through leaner architectures, optimized inference, and techniques like quantization, pruning, and distillation.
Applied AI & LLMs
Designing agentic workflows, semantic caching layers, and RAG architectures that give language models long-term recall and persistent state.
Stack
Languages
Python, TypeScript, JavaScript, SQL, C, Java
Machine Learning
PyTorch, Scikit-learn, NumPy, Pandas, Model Training & Evaluation
LLM / Applied AI
LLM APIs, Prompt Engineering, RAG, Embeddings, AI Agents, Conversational AI
AI Safety & Efficiency
Model Evaluation, Interpretability, Red-teaming, Quantization, Pruning, Distillation, Inference Optimization
Frameworks
React, Next.js, React Native, Expo
Infrastructure
Docker, Git, Linux, Cloud Deployment
Familiar
Redis, AWS, Kubernetes, WebRTC, BLE Mesh Networking
Education
B.Tech in CSE (AI & ML) at SRM AP, CGPA 8.9
Experience
Apr 2026 – Jul 2026
AI Engineer Intern at Single Core Labs - Developing applied AI products, agentic workflows, and LLM-powered applications.
Dec 2025 – Present
Founder at Zeny's Lab - Building shared infrastructure for AI, so it can grow the way humanity did: by working together, sharing knowledge, and building on each other's work.
Nov 2025 – Mar 2026
Co-Lead (Engineering) at Microsoft Student Community, SRM AP - Led engineering initiatives and mentored students in software development.
Sep 2025 – Present
Lead, AI Research Lab at Singularity Research Lab - Directed research in neural architectures and ML paradigms.
Projects
Trace
100% local activity memory engine and MCP-compatible context provider. Built with TypeScript. trace.zenyslab.com
Automail
AI-powered email automation system with Gmail OAuth and LLM APIs, serving 500+ users at peak. GitHub
WazPay
Android application using Kotlin and Accessibility Services to automate offline UPI payments through *99# USSD. GitHub
Sockets
Collaborative editor supporting real-time synchronization and peer-to-peer voice communication using WebRTC. GitHub
Scopzi
Turn messy client requests into clear scope, structured change orders, and billable work. scopzi.zenyslab.com
Buddy
Mobile AI companion using React Native, Supabase, and FCM with persistent conversational memory and real-time sync. GitHub