I build AI applications, machine learning systems, and software solutions — training models, shipping LLM agents, and turning research into production. Currently deep in Python, ML, and LLM engineering.
No traditional path, no shortcuts. I'm learning the full stack of modern AI — from linear algebra to production LLM systems — one project at a time.
I saw AI reshaping every industry and refused to sit on the sidelines. I started with Python, kept going, and never stopped.
Books, papers, docs, YouTube, and — most of all — shipping. I turn every concept into a project until it clicks.
Become a professional AI Engineer. Get hired on the strength of my work: real systems, clean code, measurable impact.
LLM agents, ML pipelines, APIs, and end-to-end AI products. I care about latency, cost, and reliability — not just demos.
A living stack. I add depth every week — from fundamentals to production systems.
Each project starts with a real problem, ends with a working system, and is documented on GitHub.
Retail traders drown in noise and lack structured, risk-aware analysis.
Multi-agent AI system that analyzes markets, manages risk, and generates trading insights. Specialized agents (research, risk, execution) coordinate through a shared plan and memory.
General chatbots forget context and can't act on your real data.
An assistant that reasons across your files, retrieves knowledge with hybrid RAG, and executes tools — calendars, search, code — through a typed function-calling layer.
End-to-end ML pipeline: feature store, gradient-boosted models, evaluation dashboards, and a REST API deployed on the cloud with monitoring.
Fine-tuned transformer on a curated domain dataset, with a lightweight inference service and a labeling loop to improve accuracy over time.
CNN-based anomaly detection with class-imbalance handling, calibrated thresholds, and a simple review UI for human-in-the-loop labeling.
Middleware that routes prompts across models by task, latency, and cost — with caching, retries, and structured logs for evaluation.
A roadmap I revisit every quarter. Every entry is backed by projects, notes, and commits.
Python mastery, data structures & algorithms, math for ML, and first ML projects. Built the habit of daily commits and public learning.
Machine learning, deep learning, and LLM engineering. Fine-tuning, RAG, agents, and serving models behind APIs.
Shipping production AI systems — reliability, evals, cost, and observability. Targeting AI Engineer roles at product-driven teams.
Commits, repos, and public work. Recruiters can verify everything on GitHub.
Multi-agent LLM system for market analysis and risk-aware trading insights.
Hybrid retrieval + tool-calling personal assistant with pgvector.
Middleware that routes prompts across LLMs by task, latency, and cost.
Production ML pipeline with feature store, evals, and monitoring.
Short essays on things I actually learned by doing — no fluff.
Planning, tools, memory, and the failure modes nobody warns you about. A postmortem on my first real LLM agent.
Attention, positional encoding, and why the architecture eats every other sequence model. Diagrams and code.
A plan for self-taught engineers: curriculum, projects, portfolio, and how to actually get hired.
Open to AI Engineer roles, freelance projects, and collaborations. Fastest reply is email.
Prefer async. I read every message.