SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Code Metal is building the leader in automated software engineering with formal verification. The company uses AI to write code while ensuring correctness through formal methods and keeping engineers in the loop. Customers include the U.S. Air Force, L3Harris, RTX, and Toshiba. Founded in 2023 with offices in Boston and San Francisco, Code Metal is backed by Accel, Salesforce Ventures, B Capital, and others.
As Staff AI Platform Engineer, you will be the technical lead of a new four-person team building the AI enablement stack that powers Code Metal's engineering teams. You'll architect and operate production systems for AI-driven code transpilation and mission planning, starting as an internal platform but built to product standard.
Your primary focus is designing, building, and operating production systems. You'll work closely with the Applied AI Research team and occasionally run experiments to validate platform decisions.
Key responsibilities include:
- Set technical direction and architecture for Code Metal's AI platform; lead the team building it. Own design docs and RFCs, guide build-vs-buy decisions, and mentor team members.
- Deploy, benchmark, and tune production inference for open-weight models on vLLM, SGLang, and TensorRT-LLM.
- Own the model gateway providing consistent auth, routing, failover, quotas, and cost attribution across self-hosted and commercial models.
- Design reusable agent harnesses and orchestration primitives for reliable, verifiable workflows.
- Build context-engineering services for memory, retrieval, and data discovery to optimize agent context and cost.
- Instrument the stack end-to-end with OpenTelemetry traces and metrics; build experiment-tracking and artifact layers for reproducibility.
- Design for productization from day one (multi-tenancy, versioned APIs, security, air-gapped deployment); partner with Applied AI Research, product teams, and DevOps.
Requirements:
- Production-grade Python and strong platform engineering fundamentals: API and service design, distributed systems, containers and Kubernetes, CI/CD, and testing.
- Shipped production agentic systems with clear understanding of failure modes and reliability improvements.
- Context engineering experience: retrieval-augmented generation, embeddings, vector or hybrid search, and agent memory, ideally over code or large technical corpora.
- Experience instrumenting services (OpenTelemetry tracing and metrics) and operating AI services against SLOs.
- Solid data science and AI research fundamentals: transformer and LLM inference mechanics, experiment design, benchmarking, and model evaluation. Working familiarity with PyTorch and Hugging Face; experience fine-tuning, evaluating, or serving language models.
- Staff-level technical leadership: owned architecture across multiple systems or teams, written design docs and RFCs, translated ambiguous internal customer needs into roadmaps, and mentored engineers.
- Typically 8+ years of software engineering experience, including 4+ years building and operating ML or LLM systems in production, with demonstrated Staff-level scope and impact. Equivalent depth valued over rigid year count.
Preferred qualifications include production experience with LLM gateways (SMG, Bifrost) or API gateways; hands-on LLM inference engine deployment with measurable gains; inference optimization knowledge (speculative decoding, prefix caching, parallelism); evaluation harnesses and experiment-tracking systems (MLflow, Weights & Biases); experience taking internal platforms to external products; and experience deploying AI systems on-prem, air-gapped, or in regulated domains like defense or aerospace.
