Machine Learning Engineer | Building efficient, scalable AI systems
🔄 Now (April 2026): Implementing papers → production. Currently implementing the memgpt paper
| Area | What I'm actually doing |
|---|---|
| Agentic Systems | Building reasoning loops + tool use |
| Performance Engineering | Making inference faster (CUDA, quantization, FlashAttention) |
| Transformer internals | From paper → PyTorch → optimized kernel |
Also exploring: RAG, SSL, multi-modal
| Repo | What it does | Status |
|---|---|---|
| search_in_terminal | CLI tool for AI answers in your terminal. Free API. | ✅ Working. Shiped April 2026. |
| Research-assistant-agent | Agentic system that downloads, reads, and summarizes research papers. | 🔄 Core loop working. Active iteration. |
| 100-days-of-cuda-kernels | Learning CUDA by writing kernels. Building toward optimized Transformer components. | 🏗️ In progress. [4] kernels so far. |
Primary: Python, PyTorch, HuggingFace, Git
Systems: C (optimization), SQL
Exploring: CUDA, Triton, vLLM
Twitter · LinkedIn · Kaggle · Portfolio · tolulopee90@gmail.com