Founding Machine Learning Engineer

Mira is building AI glasses that capture your day and feed those memories to your personal AI. We're a small team of engineers and builders in San Francisco working on the next evolution of human-computer interaction.

We've shipped our first batch of glasses and raised over $6M from General Catalyst, Naval Ravikant, Pillar VC, and more.

See the glasses in action: trymira.com

The Role

We're looking for a Founding Machine Learning Engineer to build the intelligence behind Mira. You'll own ML systems from early experiments through production, working directly with the founders and software and firmware engineers to make AI useful throughout someone's day.

This role is full-time and in-person in San Francisco.

What You'll Build

• Multimodal systems that turn audio, images, and conversations into useful memories and context.
• Memory and retrieval systems that help a personal AI find the right information at the right time.
• Evaluation datasets and experiments that measure accuracy, grounding, latency, and usefulness in real-world scenarios.
• Training and fine-tuning pipelines where adapting a model improves the product, with clear comparisons against simpler approaches.
• Fast, reliable inference across cloud and device constraints, working with the team on latency, cost, memory, and power tradeoffs.
• Production data pipelines and feedback loops that improve models while respecting user consent, privacy, and control over their data.

What We're Looking For

• 3+ years of engineering experience, including hands-on experience building and shipping ML systems.
• Strong Python skills and experience with PyTorch or a comparable ML framework.
• Solid machine learning fundamentals and experience with language, vision, or audio models.
• Experience owning the full ML development cycle: data preparation, experimentation, evaluation, deployment, and monitoring.
• Strong software engineering and systems fundamentals. You can debug the model, the data, and the surrounding application.
• Product intuition and the ability to turn an ambiguous problem into a working feature, measure the result, and iterate quickly.

Especially Relevant Experience

• Multimodal models, speech or streaming audio, computer vision, or real-time AI applications.
• Embeddings, retrieval, reranking, long-term memory, or context selection for LLMs.
• Model fine-tuning, distillation, quantization, or inference optimization.
• On-device ML, mobile or wearable products, or resource-constrained inference.
• Building consumer products or working as an early engineer at a startup.