M
Forward Deployed AI Engineer
Malleable
3 hours ago
Full-time
On-site
Brooklyn, New York, United States
Client lead, architect, and senior builder.
Full-Time Remote (US) Senior
Remote anywhere in the US. Boston preferred.
The Opportunity
Lead the work, end to end.
On any given engagement you are the technical lead, the architect, a senior builder, the manager of the Malleable engineers staffed with you, and the person who teaches the client's team to take the system over after we roll off. You are a senior engineer who can lead an architecture conversation with a VP of Engineering on Monday, ship production code midweek, review PRs from your pod, and pair with the client's team on Friday. You do this while running two or three client engagements at a time. This role suits people who want ownership of outcomes, variety across industries, and a direct hand in shaping how a small AI services firm grows up. Responsibilities What You'll Do
You'll lead engagements, lead engineers, and ship the hard parts yourself. Lead client engagements end to end
Own the technical relationship from discovery through handoff: stakeholder workshops, technical roadmap, and delivery against the business outcome. Lead the engineers staffed with you
Each engagement runs with a mix of engineers under your direction, including senior and staff peers. You set architectural direction, run PR review, and keep multiple engagements on track. Discovery and standups in Slack and Zoom, async PR review, plus 2–3 onsite trips a year for kickoffs and major milestones. Architect the systems
Agent architectures, RAG pipelines, eval frameworks, multi-agent orchestrations. Each one shaped to the client's stack, risk tolerance, and scale. Build the product
Python and FastAPI on the backend, React and Next.js on the front end. OpenAI Agents SDK and Claude Agents SDK as the primary agent frameworks. Stay hands on where it matters
You write the hard parts yourself (architecture, integrations, core agent logic) and hand the smaller tickets to engineers you are guiding through review. Stand up evals
Frameworks that track model and prompt performance over time as the application, tools, and agents evolve. Integrate with enterprise systems
MCP servers, custom agent tools, and integrations into client systems like Salesforce, Jira, SAP, internal APIs, and data warehouses. Train the client's team
Pair with their engineers, write the patterns into runbooks, and build internal champions so the system keeps moving after we are gone. Feed the practice
The components, reference architectures, and playbooks you create flow back into our toolkit so the next engagement starts further along. Requirements What We're Looking For
A senior engineer who has led delivery in client-facing settings and shipped real LLM features. Required
Experience with Python and FastAPI on the backend and React and Next.js on the front end. Production experience with OpenAI Agents SDK or Claude Agents SDK, plus you have built and maintained eval frameworks for production features. Daily use of Claude Code or Codex. The LLM is part of how you work, not a side experiment. Strong written and verbal communication, and good product instincts. You can lead a discovery with a CTO, run an exec readout, and turn a vague business problem into a scoped plan with defensible trade-offs. US Citizen or Permanent Resident. No visa sponsorship now or in the future. Strongly Preferred
You have built MCP servers or custom agent tooling, and you have shipped LLM features (GPT, Claude, Gemini) in real enterprise apps beyond chatbot wrappers. Working familiarity with LangGraph or LangChain and the Vercel AI SDK, with a clear point of view on when to use each. Cloud experience in AWS, Azure, or GCP, plus Vercel and Supabase. Comfortable with Docker, GitHub Actions, and microservice architectures. You have shipped on at least one relational database (Postgres, MS SQL Server, MySQL) and one NoSQL store (DynamoDB, Cosmos DB, Firebase). A track record of teaching or training engineers. Internal enablement programs, bootcamp instruction, conference talks, or technical writing. Experience defining and tracking post-engagement success metrics: adoption, throughput, cycle time, quality indicators, or business KPIs. East Coast home base. Our team is on the East Coast, so candidates in or near Boston or New Haven, CT are a plus for working hours overlap and the occasional in-person. Nice to Have
IaC experience (Terraform), schema migrations, secret management, and secure coding practices. Open source contributions to agent frameworks, MCP servers, or AI developer tooling. Founding engineer or early-stage startup experience where you owned 0-to-1 delivery. Why This Role Why You Might Want This
Senior leadership scope
You sit alongside the founders, with real influence on hiring, technical direction, and how we run engagements. Real outcomes, real clients
You own real outcomes for real clients, not toy demos. Variety across industries
You see a lot of different companies, stacks, and problem shapes in a year, instead of one roadmap. Founders in the work with you
Process Interview Process
Application review. We read your note, resume, and the AI system link you share. Intro call. 30 minutes with a co-founder. Take-home build. A scoped problem you do on your own time. Use any AI tools you want. We expect it. Live readout. Walk us through what you built, then answer architecture and trade-off questions on the fly without AI assistance. Final conversation and offer. Closing chat with the founders.
#J-18808-Ljbffr
On any given engagement you are the technical lead, the architect, a senior builder, the manager of the Malleable engineers staffed with you, and the person who teaches the client's team to take the system over after we roll off. You are a senior engineer who can lead an architecture conversation with a VP of Engineering on Monday, ship production code midweek, review PRs from your pod, and pair with the client's team on Friday. You do this while running two or three client engagements at a time. This role suits people who want ownership of outcomes, variety across industries, and a direct hand in shaping how a small AI services firm grows up. Responsibilities What You'll Do
You'll lead engagements, lead engineers, and ship the hard parts yourself. Lead client engagements end to end
Own the technical relationship from discovery through handoff: stakeholder workshops, technical roadmap, and delivery against the business outcome. Lead the engineers staffed with you
Each engagement runs with a mix of engineers under your direction, including senior and staff peers. You set architectural direction, run PR review, and keep multiple engagements on track. Discovery and standups in Slack and Zoom, async PR review, plus 2–3 onsite trips a year for kickoffs and major milestones. Architect the systems
Agent architectures, RAG pipelines, eval frameworks, multi-agent orchestrations. Each one shaped to the client's stack, risk tolerance, and scale. Build the product
Python and FastAPI on the backend, React and Next.js on the front end. OpenAI Agents SDK and Claude Agents SDK as the primary agent frameworks. Stay hands on where it matters
You write the hard parts yourself (architecture, integrations, core agent logic) and hand the smaller tickets to engineers you are guiding through review. Stand up evals
Frameworks that track model and prompt performance over time as the application, tools, and agents evolve. Integrate with enterprise systems
MCP servers, custom agent tools, and integrations into client systems like Salesforce, Jira, SAP, internal APIs, and data warehouses. Train the client's team
Pair with their engineers, write the patterns into runbooks, and build internal champions so the system keeps moving after we are gone. Feed the practice
The components, reference architectures, and playbooks you create flow back into our toolkit so the next engagement starts further along. Requirements What We're Looking For
A senior engineer who has led delivery in client-facing settings and shipped real LLM features. Required
Experience with Python and FastAPI on the backend and React and Next.js on the front end. Production experience with OpenAI Agents SDK or Claude Agents SDK, plus you have built and maintained eval frameworks for production features. Daily use of Claude Code or Codex. The LLM is part of how you work, not a side experiment. Strong written and verbal communication, and good product instincts. You can lead a discovery with a CTO, run an exec readout, and turn a vague business problem into a scoped plan with defensible trade-offs. US Citizen or Permanent Resident. No visa sponsorship now or in the future. Strongly Preferred
You have built MCP servers or custom agent tooling, and you have shipped LLM features (GPT, Claude, Gemini) in real enterprise apps beyond chatbot wrappers. Working familiarity with LangGraph or LangChain and the Vercel AI SDK, with a clear point of view on when to use each. Cloud experience in AWS, Azure, or GCP, plus Vercel and Supabase. Comfortable with Docker, GitHub Actions, and microservice architectures. You have shipped on at least one relational database (Postgres, MS SQL Server, MySQL) and one NoSQL store (DynamoDB, Cosmos DB, Firebase). A track record of teaching or training engineers. Internal enablement programs, bootcamp instruction, conference talks, or technical writing. Experience defining and tracking post-engagement success metrics: adoption, throughput, cycle time, quality indicators, or business KPIs. East Coast home base. Our team is on the East Coast, so candidates in or near Boston or New Haven, CT are a plus for working hours overlap and the occasional in-person. Nice to Have
IaC experience (Terraform), schema migrations, secret management, and secure coding practices. Open source contributions to agent frameworks, MCP servers, or AI developer tooling. Founding engineer or early-stage startup experience where you owned 0-to-1 delivery. Why This Role Why You Might Want This
Senior leadership scope
You sit alongside the founders, with real influence on hiring, technical direction, and how we run engagements. Real outcomes, real clients
You own real outcomes for real clients, not toy demos. Variety across industries
You see a lot of different companies, stacks, and problem shapes in a year, instead of one roadmap. Founders in the work with you
Process Interview Process
Application review. We read your note, resume, and the AI system link you share. Intro call. 30 minutes with a co-founder. Take-home build. A scoped problem you do on your own time. Use any AI tools you want. We expect it. Live readout. Walk us through what you built, then answer architecture and trade-off questions on the fly without AI assistance. Final conversation and offer. Closing chat with the founders.
#J-18808-Ljbffr