Resume & Background
Maaz Amir
AI Researcher · Self-Taught Developer · Technologist
Passionate and driven AI researcher and developer with a self-taught background spanning machine learning, blockchain, and full-stack web development. I thrive at the intersection of intelligent systems and innovative technology — building tools that push the boundaries of what's possible.
AI Research & Development Lead
Leading independent AI research focused on large language models, retrieval-augmented generation (RAG), and applied ML systems. Designed and deployed production AI pipelines integrating OpenAI, Anthropic, and open-source models into real-world applications.
- Researched and prototyped LLM-based applications including intelligent agents, document Q&A systems, and code assistants.
- Built RAG pipelines with vector stores (Pinecone, ChromaDB) to enable semantic search over large document corpora.
- Explored fine-tuning strategies for open-source models (LLaMA, Mistral) for domain-specific tasks.
Full-Stack Developer & Web3 Builder
Built and shipped full-stack web applications and decentralized applications (dApps) for clients and personal projects. Worked across the entire development lifecycle from architecture and design to deployment on cloud and decentralized infrastructure.
- Developed multiple React/Next.js applications with server-side rendering, API routes, and real-time features.
- Built smart contracts in Solidity for token systems and NFT marketplaces on Ethereum and EVM-compatible chains.
- Implemented authentication systems, REST and GraphQL APIs, and CI/CD pipelines using GitHub Actions.
Self-Directed AI & Machine Learning
Deep self-study program covering the complete machine learning stack — from mathematical foundations (linear algebra, calculus, probability) through classical ML algorithms, deep learning architectures, and modern transformer-based models. Supplemented with courses from fast.ai, DeepLearning.AI, and Stanford CS229/CS224N.
Completed over 1,000 hours of structured learning covering: supervised and unsupervised learning, neural network architectures (CNNs, RNNs, Transformers), reinforcement learning fundamentals, and modern LLM techniques including RLHF, fine-tuning, and prompt engineering.