- I am a Fullstack Software Engineer @ Cisco
- 🌱 I'm currently learning ML and AI engineering from first principles:
- Neural networks from scratch: NumPy MLP with hand-written backprop, then PyTorch
- Transformers: building a small GPT by hand (attention, sampling, KV cache)
- Reinforcement learning for LLMs: policy gradients, PPO, GRPO
- Post-training: LoRA fine-tuning and preference tuning (DPO) on small open models
- Agents & evals: tool-calling agents, retrieval, LLM-as-judge, and tracing with OpenTelemetry
- 🔨 Now building: a NumPy MLP on MNIST, with no autograd and every gradient derived by hand
- 🗺️ Everything is built in public, so follow along in my repos
- 📫 How to reach me: sean_c_benson@outlook.com
- ⚡ Fun fact: I am a marathon runner!
