Agent Engineer: A Practical Career and Portfolio Guide
Understand agent engineering across tools, state, evaluation and execution, with a portfolio exercise and scoped employer examples.
Published
Forward-deployed engineering, GTM engineering, and the roles the AI era is creating: what they are, who hires, how to get in.
A focused reading order to help you choose your next step.
Learn what Forward Deployed Engineers do, how the role works, why AI companies hire them, and when the FDE model makes business sense.
Build the engineering, product, and customer skills needed for FDE roles with portfolio projects, a 12-week roadmap, and job-search guidance.
Prepare for FDE interviews across coding, systems design, customer discovery, product judgment, ambiguity, and behavioral evaluation.
Run FDE engagements from qualification to production with clear artifacts, exit criteria, adoption measures, handoffs, and product feedback.
Compare FDEs, solutions architects, consultants, implementation engineers, and product engineers by ownership, coding depth, and outcomes.
A realistic 2026 roadmap for transitioning into an AI career: skills to learn, projects to ship, salaries to expect, and how long it actually takes.
Career field guide
FDE work connects a customer problem to software that runs in the real environment. The role combines discovery, implementation and responsibility for the outcome. Use this path to understand the job, build evidence of the skills and evaluate a team before joining it.
Ten FDE guides, with two broader guides for career transition and technical foundations.
Understand the work and the ownership boundary.
Compare the role with adjacent customer-facing jobs.
Map existing experience to a practical transition plan.
Fill gaps in the underlying technical skills.
Choose a portfolio project that demonstrates implementation.
Practice discovery, design and debugging conversations.
Follow an engagement from discovery through production.
Understand the enterprise integration context.
Decide whether the delivery model fits the problem.
Define team interfaces and recurring responsibilities.
Connect delivery measures to customer outcomes.
Review failure patterns before repeating them.
Checked September 17, 2026. Annual USD ranges from three specific US listings. These are not market averages, comparable levels or guaranteed offers. Read the linked posting for the current range and package.
| Employer and role | Location | Posted range | What the range covers |
|---|---|---|---|
| OpenAI Forward Deployed Software Engineer | San Francisco | $185,000–$325,000 | Posting labels this compensation and lists equity separately. |
| Anthropic Forward Deployed Engineer | New York / San Francisco / Seattle | $280,000–$320,000 | Posting labels this annual salary; confirm the package and level with the employer. |
| Palantir Forward Deployed AI Engineer | New York | $135,000–$200,000 | Estimated annual salary; posting excludes potential bonuses, benefits and long-term incentives from this range. |
The FDE and GTM hiring directory links directly to five employers. Compare implementation ownership, travel and post-launch responsibilities alongside pay.
Observed 2026-09-17 (UTC). 65 matching advertisements across five employer boards.
A five-employer sample of public advertisements, not a count of vacancies, people hired or the whole FDE market. Separate posting IDs can describe similar work in different locations; one posting can cover several openings. This is the first snapshot, so no growth rate is reported.
Scroll horizontally to see the full hiring table.
| Employer | Matching advertisements | Board postings inspected | Source observed (UTC) |
|---|---|---|---|
| OpenAI | 17 | 817 | |
| Anthropic | 4 | 607 | |
| Baseten | 2 | 99 | |
| Clay | 0 | 58 | |
| Palantir | 42 | 313 |
Download posting-level CSV · Snapshot and methodology data
Method fde-title-sample-v1: count unique posting IDs within each employer whose title begins with Forward Deployed Engineer, Forward Deployed Software Engineer or Forward Deployed AI Engineer (hyphens accepted). Exclude unlisted postings and titles containing manager, management, director, head, strategist, sales, intern or internship. New-graduate roles are included. Titles such as Forward Deployed Infrastructure Engineer and GTM Engineer do not match; domain or location suffixes after an eligible prefix can match. This is a title rule, not an assessment of every job's responsibilities or seniority. A failed board is unavailable, never zero; no combined total is shown for an incomplete sample.
Source dates have different meanings: Ashby reports last publication, Greenhouse first publication when supplied, and Lever a creation timestamp. Creation is not a verified publication date. Missing dates remain blank. The checked date records collection, not when a job opened. Primary and additional locations are preserved; no seniority or vacancy count is inferred from them.
Ashby uses its full-board v1 response; Greenhouse's declared total must match the returned posts; Lever pages are collected until an empty page. Response hashes, counts and request URLs are in the JSON. This is a timed observation, not an atomic snapshot across employers; changes while a board is being read can still affect it.
Missing publisher dates are left blank. The JSON records eligible advertisements and reconciled exclusion counts. Closed or removed posts leave the next approved snapshot; their disappearance does not prove a hire. Changes in title, location or board structure can change the count.
Monthly collection creates a proposed update for review. This table remains on the last approved snapshot until that review is complete.
Original practice prompts, not leaked questions or a claim about any employer’s interview loop.
Clarify the user, current workflow, costly failure and measurable acceptance criterion before proposing a model.
Collect representative failures, inspect retrieval and tool traces, separate system faults from model faults, and build a regression set.
Explain idempotency, retry limits, backpressure, observable failure states and a fallback that preserves user work.
Describe the tradeoff, who owns the exception, its expiry or migration path, and what should become a reusable product capability.
Identify irreversible actions, constrain tool permissions, preview concrete changes, and preserve an audit trail with recovery behavior.
A working artifact, acceptance results, known limitations, operational ownership and documentation someone else can use.
Understand agent engineering across tools, state, evaluation and execution, with a portfolio exercise and scoped employer examples.
Published
Map AI engineering responsibilities to a learning path, an evaluation-driven portfolio and current employer-posted compensation examples.
Published
Understand GTM engineering through actual employer roles, a practical reading path, scoped salary examples and a portfolio project.
Published · Updated
Learn what Forward Deployed Engineers do, how the role works, why AI companies hire them, and when the FDE model makes business sense.
Published · Updated
Compare FDEs, solutions architects, consultants, implementation engineers, and product engineers by ownership, coding depth, and outcomes.
Published · Updated
Use this practical decision framework to determine whether your company needs an FDE, a solutions architect, implementation help, or product work.
Published · Updated
Design an FDE team that ships customer outcomes without becoming custom services. Covers charter, structure, hiring, capacity, and career paths.
Published · Updated
Run FDE engagements from qualification to production with clear artifacts, exit criteria, adoption measures, handoffs, and product feedback.
Published · Updated
Build an FDE scorecard covering customer outcomes, delivery health, product leverage, capacity, ROI, and the metrics that create bad behavior.
Published · Updated
See how FDEs close the enterprise AI deployment gap across discovery, evals, data, integration, governance, rollout, and product learning.
Published · Updated
Build the engineering, product, and customer skills needed for FDE roles with portfolio projects, a 12-week roadmap, and job-search guidance.
Published · Updated
Prepare for FDE interviews across coding, systems design, customer discovery, product judgment, ambiguity, and behavioral evaluation.
Published · Updated
Diagnose the custom-work trap, weak qualification, permanent embedding, bad incentives, burnout, and other FDE failure modes with recovery actions.
Published · Updated
A realistic 2026 roadmap for transitioning into an AI career: skills to learn, projects to ship, salaries to expect, and how long it actually takes.
Published · Updated
Learn AI from scratch with free resources. A 90-day path from zero to building real models, with the best free courses, tools, and projects in 2026.
Published · Updated