Machine Learning Engineer, Platform
Brainco · New York City, NY
About this role
**Machine Learning Engineer, Platform** **About Brain Co.** Brain Co. builds AI-native operating systems for large, regulated institutions. We're entering our next phase of production deployments on a national scale with an elite team from Palantir, Google, Meta, and Nvidia. **The Role** As a Machine Learning Engineer on Platform, you'll build core ML capabilities that power every product we ship. You'll own capabilities end-to-end—from immediate pod needs to abstractions serving multiple teams. Your work directly impacts document extraction, foundation models, evaluation systems, and model routing across all our deployments. **Key Responsibilities** • Turn pod needs into shared platform capabilities • Own ML capabilities end-to-end in production across multiple deployments • Work at the research frontier with production stakes (LLMs, RL fine-tuning, agentic systems) • Serve project pods as true customers • Engineer for production reality: accuracy, latency, cost, and reliability • Raise the bar across the company through platform learnings **What We're Looking For** • Deep understanding of ML fundamentals: loss functions, generalization, distribution shift, evaluation • Expertise with modern AI: LLMs, agentic systems, prompting, fine-tuning, tool use, reasoning • Ability to know when fine-tuned models, VLMs, or rule engines are the right choice • Platform instinct: spotting general capabilities in specific requests • Comfort building things that have never existed before • Treat frontier models as engineered components, not magic **Problems You'll Work On** • Agents as shared capabilities across products • Foundation models for construction documents • Model routing for optimal cost, latency, and performance • Unified evaluation systems across all use cases • Composite AI systems and credit assignment • Continuous improvement machinery for production systems
Listing freshness
CronJobs last confirmed this listing 1h ago. If its source stops confirming the opening for seven days, this page is removed from active inventory.