Maxime Rivest
Bio
I am an educator and developer dedicated to crafting tools and content that make LLMs more accessible and powerful for everyone. I am a core contributor to DSPy, and I've built numerous open-source Python libraries to advance the ecosystem, including attachments, functai, ovllm, dspy-lm-auth, dspy-template-adapter, and mcp2py. When I'm not coding, you'll find me spending my weekends farming or tinkering away on my next open-source project.
Session (1)
Maxime Rivest is a prolific DSPy practitioner and open-source toolmaker. With a background as a Lead Data Analyst / AI optimization engineer at Elsevier, he builds widely-adopted Python libraries like Attachments, mcp2py, and FunctAI, combining production-scale LLM experience with framework-level contributions on top of DSPy. Attending his session offers direct insight into practical DSPy optimization, MCP-as-Python-modules, and the ergonomics of building real AI systems.
Recent talks (2)
GitHub
@MaximeRivest- MaximeRivest/attachments★ 360Easiest way to give context to LLMs — transforms any file into images+text for LLM context with 2 lines of Python
- MaximeRivest/maivi★ 279Real-time voice-to-text local on CPU, better than Whisper, with hotkey support
- MaximeRivest/mcp2py★ 252Turn any MCP server into a Python module with zero configuration
- MaximeRivest/functai★ 85Vanilla-Python function-style API for building AI agents with DSPy
- MaximeRivest/funnydspy★ 40Vanilla-Python ergonomics on top of DSPy
Recent writing (6)
- The Signature Duality: Separating Intent from Mechanism in DSPy · blog · Feb 2026
- DSPy sounds neat, but is it reliable? · blog · Sep 2025
- Hacking DSPy into doing Automatic System Prompt Optimization · blog · Jul 2025
- DSPy Optimizers – Parameter Structure Analysis (by deep research) · blog · Jul 2025
- How to Build an Automatically Branching Chat with DSPy · blog · Jul 2025
- A Simple Introduction to DSPy · blog · Jul 2025
Public activity researched automatically · as of Jun 2026