Forward Deployed Engineer
About The Position
This is the person we send into a client when the problem is not yet a spec. You sit with their operators, finance leads, and analysts, watch how work actually gets done, and turn messy real-world processes into agentic AI workflows that hold up in production. You are dual-fluent by design. In the morning you map a claims-handling or order-to-cash process with a business owner who has never written a line of code. In the afternoon you wire that process into an agent that calls their CRM, queries their warehouse, and routes exceptions to the right human. You translate in both directions, turning vague business pain into concrete technical scope, and turning technical constraints into options a non-technical stakeholder can decide on. The role is a balanced bridge. Roughly half your time is discovery, facilitation, and trust-building with client teams. The other half is hands-on engineering: building the agent logic, the integrations, and the evaluation that proves it works. You will not stop at a prototype. You ship the workflow, instrument it, and hand over something the client can run and extend.
Requirements
- 3 or more years of software engineering experience building systems that run in production, plus the judgment to know what production-ready means
- Hands-on experience building with LLMs: prompting, function calling, structured output, retrieval, and at least one agent or multi-step workflow you took past the prototype stage
- Strong integration skills, you have connected real systems through REST APIs, webhooks, SDKs, and databases, and dealt with auth, pagination, rate limits, and flaky upstreams
- Proficiency in Python, comfort with async patterns, API design, and the data-pipeline reliability that agentic workflows live or die on
- Working knowledge of SQL and data systems, enough to pull what a workflow needs from a client's warehouse or operational database
- Real client-facing or cross-functional skill: you can sit with a non-technical stakeholder, draw out how their process works, and earn their trust without talking down to them
- Strong written and verbal communication, you can explain why an agent took a wrong turn, or why a manual step should stay manual, to a business owner without losing them
- A bias toward shipping, you would rather get a working workflow in front of users this week than perfect an architecture diagram
- Production experience with agent frameworks (LangGraph, AutoGen, CrewAI) or building custom orchestration
- Experience with RAG systems: chunking, embeddings, vector stores (Pinecone, Weaviate, Qdrant, pgvector), and re-ranking
Nice To Haves
- Familiarity with LLM evaluation: task-specific metrics, LLM-as-judge, and regression test suites
- Background in process mapping, solutions engineering, forward-deployed work, or technical consulting across concurrent engagements
- Experience with workflow and integration platforms (Temporal, Airflow, Zapier, n8n) or iPaaS tooling
- Exposure to a major cloud and its AI services (AWS Bedrock or SageMaker, GCP Vertex AI, Azure ML)
Responsibilities
- Embed with client teams to understand how work actually happens, shadowing operators and mapping processes that live in people's heads, spreadsheets, and a dozen disconnected tools
- Break down complex business processes into agentic workflows: decompose a goal into steps, decide what an agent automates versus where a human stays in the loop, and design the tool-use and decision logic to match
- Build production-grade agentic workflows using frameworks like LangGraph and LangChain, including tool calling, memory, routing, retries, and clean failure recovery
- Build integrations to the data and operational systems that workflows depend on: CRMs, ERPs, ticketing systems, data warehouses, internal APIs, and document stores, handling auth, rate limits, and reliability
- Enable core platform capabilities that the whole engagement reuses: retrieval over client documents, structured output and function calling, prompt versioning, logging, and observability
- Design evaluation and guardrails that make a workflow trustworthy in an enterprise setting: task-specific evals, regression checks, PII handling, and human review steps where the stakes demand them
- Run working sessions and demos that keep business stakeholders bought in, translating technical tradeoffs into clear decisions about cost, risk, and timeline
- Document what you build and hand it over cleanly, so client teams can operate and extend the workflows after the engagement
- Bring patterns back to BOD, turning what worked on one engagement into reusable building blocks for the next
Benefits
- Competitive compensation with performance bonuses
- Problems that matter, real client processes with real constraints, not greenfield sandboxes
- A front-row seat to how a dozen industries actually run, and the chance to automate the parts that slow them down
- Direct access to modern AI infrastructure: AWS, GCP, Azure, Databricks, Snowflake, and the open-source ecosystem
- Builder culture where engineers lead and ship, with no bureaucracy between you and the work
- Professional development budget and certification support
- A team that will push back on your design decisions and make them better
- Unlimited Paid Time Off (PTO)
- Paid parental/bereavement leave
Stand Out From the Crowd
Upload your resume and get instant feedback on how well it matches this job.
What This Job Offers
Career Level
Senior
Education Level
No Education Listed
