Jobs · Massachusetts

Staff AI Platform Engineer

Form Energy · Somerville, MA · 4 wk ago
Hybrid$129k–$185k/yrFull-time

About the role

Form Energy is an American manufacturing and energy technology company revolutionizing energy storage with cost-effective, multi-day technology designed to keep the electric grid secure and reliable, even during extended periods of stress. The company has been named by TIME as a “Best Invention”, by MIT Technology Review as a “Top Climate Tech Company To Watch”, and by Fast Company as “One of the Next Big Things In Tech”. With signed contracts with leading electric utilities across the United States and production of iron-air batteries underway at a high-volume manufacturing facility in West Virginia, Form Energy is scaling rapidly.

Form Energy is hiring a Staff AI Platform Engineer to build the shared platform that lets people across the company build and run AI-powered tools safely and quickly. As a hands-on senior individual contributor, you will help design the application runtime, knowledge systems, agent infrastructure, and governance tooling that teams would otherwise build separately, and you will help set the patterns other engineers follow. Your work spans product R&D, design, manufacturing process engineering, product costing, and root cause analysis, drawing on data from battery test stands, factory manufacturing robots, deployed projects in the field, and the documents, slide decks, transcripts, and conversations teams produce daily. You will sit on the AI & Data Platform team, working with data engineers, software engineers, Information Security, and the scientists, engineers, and operators who use what you build.

This is a hybrid role requiring onsite work from the Berkeley, CA, Somerville, MA, or Weirton, WV office 3+ days per week. Relocation assistance is available.

Responsibilities

  • Build and operate the internal application platform that lets employees create and deploy AI-generated applications on curated backend and frontend frameworks, with starter templates, design guidance, authentication, and secure deployment built in, so app authors own only business logic, user flows, and UI.
  • Give those applications governed access to data, including connectivity to the data warehouse and a shared Postgres cluster for application state that is automatically integrated into the data platform.
  • Build and operate a central knowledge platform over unstructured content (documents, slide decks, meeting transcripts, Slack) and structured data from battery test stands, manufacturing robots, and deployed systems, pairing document retrieval with query access to structured data through semantic models and text-to-SQL. Own its retrieval quality through semantic and hybrid search, re-ranking, chunking, and classification, with source-level permissions and evaluation sets that show whether a change helps.
  • Design and deploy the shared MCP infrastructure, including a gateway layer (such as agentgateway) for authentication, authorization, rate limiting, and audit logging, and build MCP servers for applications and services across the software org and for selected external tools.
  • Own model selection and cost: evaluate open-source and open-weight models against frontier models on internal tasks, decide where each fits, and reduce token usage through context management, caching, and model routing.
  • Build the agent orchestration layer, including multi-agent workflows and long-running loops with checkpoints, human review gates, and automated evaluation. Work with data engineers to decide when work should run as a deterministic, scheduled workflow and when it needs a graph or state machine in which an agent chooses among bounded steps.
  • Partner with product R&D, battery design, manufacturing process engineering, product costing, and root cause analysis teams to find repeatable work and automate it, including agent workflows whose output informs critical design and process decisions.
  • Instrument the platform with tracing, evaluation, and usage and cost attribution by team, and stand up the AWS infrastructure and infrastructure-as-code (Pulumi and/or Terraform) it runs on.
  • Help stand up a company-wide skills and plugin marketplace that vets contributions against AI policy and scans them for security vulnerabilities with open-source and commercial tools.
  • Advise on AI governance, recommending changes that make the policy specific to platform needs and automating its execution, such as approvals for new tools.
  • Raise the engineering bar through design docs and RFCs, code review, mentorship, and technical influence that reaches beyond the immediate team.

Requirements

  • Bachelor's degree in computer science, software engineering, computer engineering, or a related field plus 7+ years of software or data engineering experience, including building and operating production platforms or services that other engineers depend on.
  • Strong production programming in Python and Go, with experience building APIs and backend services.
  • Hands-on experience building LLM-powered systems in production, including retrieval-augmented generation, embeddings, vector and hybrid search, re-ranking, and structured-data access through text-to-SQL or semantic layers.
  • Experience with agent frameworks, tool calling, and the Model Context Protocol, and with workflow orchestration or state-machine frameworks such as LangGraph, Temporal, or Dagster.
  • Experience building evaluation sets and harnesses for LLM systems and measuring quality, latency, and cost across model providers, including open-weight models.
  • Experience with PostgreSQL and a cloud data warehouse, plus AWS, containers, CI/CD, and infrastructure-as-code with Pulumi and/or Terraform.
  • Working knowledge of platform and AI security: authentication and authorization (OAuth/OIDC), secrets management, sandboxed execution, tenant isolation, supply-chain scanning, and risks specific to LLM tools such as prompt injection and data exfiltration.
  • A track record of building internal platforms that other teams adopt, and of independently leading the design of novel solutions through technical design docs and RFCs.
  • Excellent communication and organizational skills, including with scientists, engineers, and operators who aren't software developers, and comfort participating in an on-call rotation.

Skills

  • React and design-system experience
  • Snowflake or Databricks
  • Serving open-weight models (for example with vLLM)
  • API or AI gateway experience
  • AI governance or compliance frameworks (SOC 2, ISO 27001)
  • Background in battery, energy, manufacturing, or other physical-systems data

Benefits

  • 100% of medical, dental, and vision premiums covered for full-time employees (80% for dependents), effective day one
  • At least 12 weeks of paid leave for new parents (up to 20 weeks for birthing parents)
  • Generous vacation policies
  • Competitive salaries and stock options

Pay

Compensation range: $128,660 – $184,966

Schedule

Hybrid role requiring onsite work 3+ days per week from the Berkeley, CA, Somerville, MA, or Weirton, WV office. Relocation assistance available.

Similar jobs