I'm an Open Source Software Developer at IBM working in the AI inference space. I've contributed to vLLM and built a vLLM plugin for hardware enablement on the IBM Spyre Accelerator, and before that I maintained KServe/ModelMesh for model serving.
I also do developer advocacy work: 10+ international conference talks on LLM inference and model serving, mentoring new contributors, and writing blogs and tutorials.
My background includes a MASc spanning reinforcement learning, virtual reality, and autonomous vehicle research, and time spent in retail data science.
| Project | Role | What I did |
|---|---|---|
| IBM Granite Community | Maintainer, 2026–now | Workshop recipes per model release, CI build infrastructure, new models on Replicate and Ollama |
| vLLM-Spyre | Maintainer, 2025–2026 | FP8 quantization, sliding window attention, tensor-parallel tests, CI/CD with model caching, docs |
| vLLM | Contributor, 2024–2025 | Bug fixes, frontend features, CI linting, docs |
| KServe ModelMesh | Maintainer, 2022–2024 | Multi-runtime serving (Triton, TorchServe, MLServer), gRPC/dependency updates, Grafana dashboards |
I've also contributed to InstructLab, Caikit, and Machine Learning Exchange.
- Fast Inference, Furious Scaling, Open Source Summit NA 2025
- Leveraging Open Source AI Securely and Privately, Devopsdays Geneva 2025
- All talks & workshops →