About Liquid AI
Spun out of MIT CSAIL, Liquid AI builds general-purpose AI systems that run efficiently across deployment targets, from data center accelerators to on-device hardware, ensuring low latency, minimal memory usage, privacy, and reliability. They partner with enterprises across consumer electronics, automotive, life sciences, and financial services and are scaling rapidly.
The Opportunity
This role focuses on bridging the imagination gap for developers regarding the capabilities of fine-tuned small models running on various devices. You will be Liquid's technical voice on platforms like Hugging Face, GitHub, Discord, and at key edge AI developer events, transforming individual troubleshooting into comprehensive cookbooks, reference applications, and community content. This is a builder role emphasizing shipping working solutions and taking ownership.
What We're Looking For
We need someone who is a:
- Builder: Prioritizes shipping working reference apps over strategy memos and measures success by developer adoption.
- Community native: Enjoys public problem-solving, is comfortable with being wrong, and possesses strong opinions on the small-model and edge-AI ecosystem.
- Honest technical communicator: Conducts fair benchmarks, avoids overselling, and dislikes vague or AI-generated content.
- In-person energizer: Thrives at hackathons, workshops, hallway-track demos, and conferences where builders gather.
- End-to-end owner: Takes initiative in ambiguous situations, converting one-off problems into reusable resources without being prompted.
The Work
- Be the technical voice in the community: Engage where LFM developers are (Hugging Face, GitHub, Discord, X), answer technical questions, maintain Liquid's presence on the Hub, and highlight notable community fine-tunes and edge deployments.
- Show up in person: Organize and host hackathons, build nights, and technical workshops at Liquid offices and partner venues. Represent Liquid at essential small-model and edge AI conferences, submitting talks, managing booths, and demonstrating in the hallway track. Collaborate with hackathon organizers to promote LFMs.
- Build the on-device and adaptation story: Ship reference applications showcasing LFMs on real edge hardware (iOS, Android, browser, Jetson, NPUs). Maintain integration recipes for common inference stacks (llama.cpp, MLX, ONNX, executorch, LEAP SDK). Create cookbooks guiding developers from base model to fine-tuned, quantized, and deployed.
- Create content that earns trust: Produce deeply technical blog posts, demos, and honest benchmarks against models like Qwen, Gemma, Phi, and SmolLM, focusing on first-principles and concrete examples. Convert successful in-person workshop material into evergreen online content.
- Close the loop: Document developer pain points (chat templates, tokenizers, deployment issues) and integrate this feedback into roadmap discussions with the model and platform teams.
Must-have Desired Experience
- Proven technical expertise: Hands-on experience with LLMs, including model fine-tuning (LoRA, QLoRA, full fine-tuning, distillation), tokenizer debugging, and a track record of shipping to production or active community environments.
- Fluency with the modern AI stack: Deep familiarity with PyTorch, Hugging Face (Transformers, PEFT, Datasets), and model serving frameworks (llama.cpp, MLX, vLLM, ONNX Runtime, or TGI), along with an understanding of quantization tradeoffs (GGUF, AWQ, GPTQ, INT8/INT4).
- Efficient model specialization: Experience with on-device deployment (iOS, Android, embedded) or specialized work within the efficient-model ecosystem (Phi, Gemma, Qwen, SmolLM, or distilled architectures).
Nice-to-have
- Prior developer relations, developer advocacy, or community engineering experience.
- A track record of public technical communication: conference talks, widely read technical writing, or a following among ML and edge developers.
- Active open-source contributions, especially to the inference or efficient-model tooling ecosystem (llama.cpp, MLX, ONNX, Hugging Face libraries).
- Experience organizing or running hackathons, workshops, or developer events.