Vibe Engines — Interview Prep

The Commercial Layer for FDEs

Sat, 01 Aug 2026 00:00:00 GMT

The part nobody teaches: reading a SOW before you are bound by it, the six clauses that cause all the pain, four pricing models and what each does to your week, telling a customer need from a rep’s quota, the services-trap arithmetic, and the renewal that starts nine months early.

Applied AI Engineer, Explained

Sat, 01 Aug 2026 00:00:00 GMT

“Applied” does not mean applied research — it means the model meets a real customer’s systems and you are standing where that happens. The published 40/30/30 split, how the title differs from AI engineer and ML engineer, the five badge variants in the family, and the two questions that identify any of them.

FDE vs FDSE vs Deployment Strategist

Sat, 01 Aug 2026 00:00:00 GMT

Palantir split forward-deployed work into an engineering ladder and a strategy ladder; most companies copied the title and not the split, which is why one word now means two jobs. What each role owns, how the loops diverge, and the career mistake this comparison exists to prevent.

AI Solutions Engineer

Sat, 01 Aug 2026 00:00:00 GMT

The engineer on the other side of the signature. Roughly a third of FDE-family postings are pre-sales-shaped: what that means across the deal cycle, how OTE changes which customers you spend time on, why “how accurate is it?” must not be answered with a number, and the four questions to ask before accepting.

The FDE Canon

Sat, 01 Aug 2026 00:00:00 GMT

Ten sources that actually explain the role — first-hand accounts, job descriptions read as documents, the commercial argument, and the market data — each with what to take from it and what it is silent about. Read in this order, and note honestly what is missing: almost nothing good exists on identity, security review, the commercial layer or evals.

Stakeholder Politics Lab

Sat, 01 Aug 2026 00:00:00 GMT

Deployments die on org charts, not architecture. Six people at one customer, each identified by a single sentence — classify them as economic buyer, champion, end user, blocker or data gatekeeper, and expect the titles to mislead you at least twice. Then handle the three situations that actually end projects: the gatekeeper who never says no but never grants access, the team lead whose people you are automating, and the champion who resigns in month five — while a coalition meter tracks who is still with you.

Domain Ramp Sprint

Sat, 01 Aug 2026 00:00:00 GMT

The FDE meta-skill, drilled on three real verticals — insurance claims, hospital revenue cycle, or legal discovery. Learn the vocabulary that turns out to be codes (status 7, CO-45, TAR), reconstruct the workflow people actually follow rather than the one on the diagram, then pass expert gates including the one that asks which rule lives only in a practitioner's head. Scored on whether you could hold a conversation with them on Monday.

Deep Clone, Constraint by Constraint

Sat, 01 Aug 2026 00:00:00 GMT

The incremental round on a problem where stage three genuinely breaks the obvious design: copy an object, then nested structures, then one containing a cycle, then one where two keys must still share the same clone. A visited set stops the crash and fails the last stage. Solve it in Python or TypeScript, with hidden tests.

Rate Limiter, Constraint by Constraint

Sat, 01 Aug 2026 00:00:00 GMT

A third incremental round, where stage two invalidates stage one outright: a fixed-window counter becomes a sliding window, then per-key, then gains a burst allowance. Tests whether you can say “that counter has to go” calmly and refactor. Solve it in Python or TypeScript, with hidden tests.

The 95% Number, Examined

Sat, 01 Aug 2026 00:00:00 GMT

The most-quoted statistic in enterprise AI, read carefully: roughly 95% of pilots reportedly produced no MEASURABLE profit-and-loss impact. Two words carry the sentence. "Measurable" means attributable financial effect — not accuracy, not satisfaction — and "pilot" means a bounded trial, which selects for projects that end before an accounting period closes. Three separate failures hide inside the one number: MEASUREMENT failure (it worked and no baseline exists, because nobody captured the before-picture in week one and it is unrecoverable afterwards), ATTRIBUTION failure (something improved and three other things changed the same quarter, so finance will not credit yours), and ACTUAL failure (the workflow was wrong or nobody adopted it). The first two are why the forward deployed role exists — both are solved by work before and after the model. Sourced, confidence-labelled (reported via press coverage, not a public methodology), and paired with the a16z argument it mirrors.

Trading Margin for Moat

Sat, 01 Aug 2026 00:00:00 GMT

Why a software company would deliberately hire expensive engineers to do customer work. The services-led-growth thesis: spend gross margin on deployment depth because the resulting integration is hard to displace and the outcome is provable — the margin hit is an acquisition cost for defensibility, not an inefficiency. The arithmetic that decides whether it holds, as a worked napkin example: an FDE at ~$300K loaded doing 3 deployments a year costs ~$100K per deployment, which is 67% of a $150K ACV and 17% of a $600K one — and REUSE is the only term that improves over time (0.6x, then 0.4x). The failure mode built into the thesis is paying the margin and not receiving the moat, which arrives one reasonable exception at a time. Diagnostic: does deployment N take measurably less time than N−1? Plus what the trade means for your career and the two questions to ask an employer.

1,000 FDE Jobs, Analysed

Sat, 01 Aug 2026 00:00:00 GMT

Reading a thousand job postings beats reading a thousand opinions. One title covers at least THREE distinct jobs — builder (few customers, deep, milestone-driven), pre-sales (many customers, shallow, quarter-driven, often with quota), and internal/platform — and roughly a third of postings using the FDE title are the pre-sales shape, which the posting rarely makes explicit. A separate census counted 1,206 strict-definition postings across 669 companies at a ~$185K median posted base. What every posting asks for: strong general engineering, LLM app patterns, integration reality, deployment where you do not own the cloud, and increasingly MCP servers as deliverables — with EVALUATION the most-cited hard skill and the least taught anywhere. How to read a posting properly, the two questions that settle which species it is, and the honest caveat that posting counts measure demand signals rather than filled roles.

FDE Comp Explorer

Sat, 01 Aug 2026 00:00:00 GMT

Forward Deployed Engineer compensation, decomposed. Pick an employer tier (frontier lab, Palantir, AI infra, vertical startup, big tech, defence), a level and a market, and see the base/equity/bonus split rather than one headline number — plus the reported band, the confidence label on the underlying source, and the six questions that change what any offer actually means. A transparent multiplier model with its anchors and sources stated, not a scraped table.

Scoping Question Log

Sat, 01 Aug 2026 00:00:00 GMT

A drill for the skill every discovery call actually tests. Two underspecified customer asks, twelve candidate questions each, and a budget of six — spend them, see what the customer really says, and get each question scored on whether the answer changes what you build. Ends with a copyable scoping doc, a coverage map across the six axes, and the thing that was really going on that you either found or missed.

What Is a Forward Deployed Engineer?

Fri, 31 Jul 2026 00:00:00 GMT

The pillar definition, with sources: an FDE owns the full arc of a deployment inside the customer\u2019s environment. The three species of the role (builder / pre-sales-shaped / internal), why postings grew 729% in a year, who hires, what the bar really is, and whether the title survives its own hype.

FDE vs Every Adjacent Role

Fri, 31 Jul 2026 00:00:00 GMT

One master table placing the Forward Deployed Engineer against eleven adjacent roles \u2014 software engineer, AI engineer, solutions engineer, architect, consultant, PM, DevOps, data engineer, SRE, TAM, customer success \u2014 on environment, ownership and comp shape, then the four deep comparisons and a four-check test for reading any job ad.

FDE vs Software Engineer

Fri, 31 Jul 2026 00:00:00 GMT

Same code, different physics. The ownership asymmetry that rewrites everything, two calendars side by side, why FDE comp is higher but far wider, what transfers from product engineering and what has to be built new, and honest checklists for switching in either direction.

FDE vs AI Engineer

Fri, 31 Jul 2026 00:00:00 GMT

Identical toolbox \u2014 RAG, agents, tool schemas, evals \u2014 but the AI engineer\u2019s hard problem is the model and the FDE\u2019s is the organisation around it. Where Applied AI Engineer and Agent Engineer land, negotiated evals as the skill that separates them, and how comp, ceiling and exits differ.

FDE vs Solutions Engineer & Architect

Fri, 31 Jul 2026 00:00:00 GMT

One question separates all three: after signature, who is accountable for what runs in production? Includes the 30-second test that reveals a relabelled pre-sales posting, one enterprise deal followed phase by phase, and the production-engineering gap to close if you are moving over from pre-sales.

FDE vs Consultant

Fri, 31 Jul 2026 00:00:00 GMT

Is forward deployed engineering just consulting with better branding? Three things genuinely separate them \u2014 running software, a product feedback loop, and equity \u2014 and each can quietly disappear. The services trap, four interview questions that detect it, and what FDEs should steal from consultants anyway.

Forward Deployed Engineer Salary, Explained

Fri, 31 Jul 2026 00:00:00 GMT

Why the public numbers differ by seven times, reconciled: base versus total comp, six employer tiers from frontier labs to defense integrators, the level ladder and what earns each rung, six questions for reading an offer, and the equity structure that makes most of the package an estimate about the future.

How to Become a Forward Deployed Engineer

Fri, 31 Jul 2026 00:00:00 GMT

The concrete route in: the real bar (staff-level judgement, not SWE-lite), six skills in dependency order, the one portfolio artefact that beats everything else, a 12-week plan, what to do coming from support / consulting / no degree, and an honest verdict on paid FDE courses and certifications.

Is Forward Deployed Engineering Worth It?

Fri, 31 Jul 2026 00:00:00 GMT

The honest version. Five named costs \u2014 coding atrophy, travel, accountability without control, the services trap, and a specific kind of burnout \u2014 each with the version of the job where it does not happen. Then the exits ranked by how naturally they follow, and eight questions that buy you out of the costs.

A Day in the Life of an FDE

Fri, 31 Jul 2026 00:00:00 GMT

Hour by hour, built from the published first-hand accounts: a customer-site day, a remote build day, and a launch-week day \u2014 plus how the 40/30/30 split actually swings across discovery, prototype, hardening and handoff phases of a single deployment.

FDE Interview Prep: The 9 Rounds

Fri, 31 Jul 2026 00:00:00 GMT

The complete Forward Deployed Engineer interview guide \u2014 decomposition, learning round, incremental coding, debugging/re-engineering, take-home + defence, client roleplay, values, customer-flavoured system design and behavioural ownership. What each scores, the five signals every company screens for, a four-week plan, and an interactive drill for every round.

The Palantir FDSE Interview

Fri, 31 Jul 2026 00:00:00 GMT

The loop that every other company copied: the online assessment, the panel drawn from a pool of five round types, and the decomposition round minute by minute \u2014 what strong candidates do in the first five minutes versus what fails. Plus the FDSE / Deployment Strategist split and a two-week prep plan.

The OpenAI FDE Interview

Fri, 31 Jul 2026 00:00:00 GMT

How the reported OpenAI forward deployed engineer loop works: the multi-hour API take-home, the required video walkthrough where it is actually graded, and the case/empathy rounds that carry roughly half the evaluation \u2014 with six habits that win the take-home.

The Anthropic Applied AI / FDE Interview

Fri, 31 Jul 2026 00:00:00 GMT

The most agent-shaped FDE role at a frontier lab: the escalating-constraint coding assessment, the deployment case where negotiated evals win it, and the values round that candidate reports call the highest-failure stage \u2014 plus exactly how to prepare for it.

The Sierra Agent Engineer Interview

Fri, 31 Jul 2026 00:00:00 GMT

Live incremental coding, a support agent you must build without an agent framework, and a debugging onsite with planted bugs. Why the no-framework constraint is the cleanest signal-extraction device in any published FDE loop \u2014 and the six decisions frameworks hide from you.

The Distyl AI Interview (AI Allowed)

Fri, 31 Jul 2026 00:00:00 GMT

One of the few companies that publishes its interview philosophy: use AI in the take-home, then defend everything it produced. The three anti-patterns they reject \u2014 Autopilot, Default and Demo \u2014 and a phase-by-phase split of what stays human when AI is permitted.

The FDE Take-Home Playbook

Fri, 31 Jul 2026 00:00:00 GMT

Take-homes are the centre of gravity in modern FDE loops and they are graded on judgement, not volume. The five-phase method with time budgets, what graders actually weight, a deliberately ambiguous five-hour practice spec, a 30-point self-scoring rubric, and how to prepare the video defence.

The Decomposition Round, Simulated

Fri, 31 Jul 2026 00:00:00 GMT

Don't read about the FDE decomposition round \u2014 sit it. Pick a case (taxi fleet, emergency response, fraud unification, marketplace) and get scored on the five clarifying questions you ask before the propose button unlocks, then on which sub-problems you attack and in what order, then on the two things you explicitly refuse to build. Premature solutioning scores negative, exactly like the real round. Ends with an interviewer verdict and a band. The signature Palantir-style round, interactive nowhere else.

Client Roleplay Simulator

Fri, 31 Jul 2026 00:00:00 GMT

The four conversations that actually decide deployments, as branching dialogue: your demo breaks in front of their CTO, a stakeholder insists on the wrong feature, security will not release the data, and a manager whose team gets smaller keeps raising objections. Every reply moves a live trust meter and changes what they say next. Ends with a scored verdict against the four signals FDE interviewers screen for \u2014 the round every guide tells you to rehearse with a friend.

The Learning Round, Simulated

Fri, 31 Jul 2026 00:00:00 GMT

Learn a library that does not exist. You get the internal docs for flowpkg \u2014 a package installer with dependency stages, a barrier between them, ordered post-install hooks and idempotent installs \u2014 in three unlocking sections with comprehension gates. Two rules are deliberately buried, and the last question can only be answered by combining rules from different sections. Tests reading strategy, not recall, exactly like the real round.

Re-Engineer a Legacy ETL

Fri, 31 Jul 2026 00:00:00 GMT

You inherit an aggregation nobody can explain and the totals are wrong. Three planted defects \u2014 an exclude-list where the spec wants an include-list, a dedupe key missing a field, and pre-seeded customers that should never appear. The FDE re-engineering round. Solve it in Python or TypeScript, with hidden tests.

Repair the Spread Traversal

Fri, 31 Jul 2026 00:00:00 GMT

Inherited code computes how far something spreads through a contact graph, and the answers are wrong in a way that is invisible on small inputs. A one-way adjacency map, a frontier with no dedupe, and a seen-set updated one step too late. Solve it in Python or TypeScript, with hidden tests.

LRU Cache, Constraint by Constraint

Fri, 31 Jul 2026 00:00:00 GMT

The incremental-coding round, packaged: a cache that becomes bounded, then least-recently-used, then instrumented \u2014 each stage invalidating the shape of the last. Tests whether your first version can absorb the next requirement. Solve it in Python or TypeScript, with hidden tests.

The AI-Era Data Engineer

Mon, 27 Jul 2026 00:00:00 GMT

How the data-engineer job changes in the AI era. SQL, pipeline glue and connector code get cheap; data modeling, quality and contracts, cost, lineage and the semantic layer become the moat — plus a new high-demand specialty: building the data pipelines that feed AI (RAG ingestion, embeddings, eval datasets). Five AI-augmented workflows, a judgment exercise, and a 90-day plan.

The AI-Era QA Engineer

Mon, 27 Jul 2026 00:00:00 GMT

How the QA / test-engineer job changes in the AI era. Writing test scripts, selectors and boilerplate cases gets cheap; test strategy, exploratory testing, risk-based judgment, and the new discipline of testing AI and nondeterministic systems (evals, not assertEquals) become the moat. Five AI-augmented workflows, a judgment exercise, and a 90-day plan.

The AI-Era Engineering Manager

Mon, 27 Jul 2026 00:00:00 GMT

How the engineering-manager job changes in the AI era. Status reports, summaries and first-draft reviews get cheap; people leadership, technical judgment, hiring and setting direction become the moat — plus a new mandate: leading AI-augmented teams and measuring output, not activity. Five AI-augmented workflows, a judgment exercise, and a 90-day plan.

The New Grad in the AI Era

Mon, 27 Jul 2026 00:00:00 GMT

A straight answer to “AI is killing junior roles — what do I even learn?” The junior grind (boilerplate, glue, first-draft tests) is being automated, so the bar shifts to fundamentals, judgment, and working WITH AI earlier — and you need fundamentals more, not less. What to learn now, five moves for your first months, a judgment exercise, and a 90-day plan.

AI for Support Engineers

Mon, 27 Jul 2026 00:00:00 GMT

Support automation is the #1 deployed AI use case, so this role is changing fast. What AI automates (tier-1 tickets, FAQs, routing) versus what becomes valuable (hard troubleshooting, escalation judgment, empathy, and improving the AI itself), the new support stack, and the high-value pivots — agent-ops, AI-support engineering, and forward-deployed engineering. Five moves, a judgment exercise, and a plan.

AI for Data Analysts

Mon, 27 Jul 2026 00:00:00 GMT

Text-to-SQL and auto-dashboards are automating the mechanical half of analysis. What AI automates (writing SQL, basic dashboards, pulling numbers) versus what becomes valuable (asking the right question, judging whether data can be trusted, interpreting results in business context, and defining what the metrics mean), plus the analytics-engineering escape hatch. Five moves, a judgment exercise, and a plan.

AI for Technical Writers

Mon, 27 Jul 2026 00:00:00 GMT

AI drafts docs, generates API reference from code, and answers user questions directly — so fewer people read the docs at all. But the twist most miss: your documentation is now the context layer AI assistants, agents, and RAG systems read from, and an AI that reads wrong or unstructured docs gives confident wrong answers at scale. What AI automates (first drafts, boilerplate, formatting) versus what becomes valuable (information architecture, docs-as-context, docs-as-evals, owning the source of truth), plus the pivots — docs/DX engineering, developer relations, and AI knowledge engineering. Five moves, a judgment exercise, and a plan.

AI for Business Analysts

Mon, 27 Jul 2026 00:00:00 GMT

AI drafts requirements docs, turns meeting notes into user stories, summarizes stakeholder calls, and writes the SQL for your reports. But the hard part of business analysis was never the writing — it was asking the right questions, resolving conflicting requirements, and deciding what is actually worth building. What AI automates (documentation, first-draft stories, process diagrams, basic queries) versus what becomes valuable (problem definition, stakeholder judgment, process redesign, validating that requirements match reality), plus the pivots — analytics engineer, AI product manager, and product owner. Five moves, a judgment exercise, and a plan.

AI for Technical Program Managers

Mon, 27 Jul 2026 00:00:00 GMT

AI writes your status reports, summarizes the meeting, updates the tracker, and drafts the project plan — the reporting layer of program management is being automated. But the job was never the reports; it was driving cross-team execution, killing the risks that matter, and forcing decisions when senior people disagree. What AI automates (status, notes, trackers, first-draft plans) versus what becomes valuable (cross-org coordination, risk judgment, technical depth to challenge estimates, unblocking), how the TPM role differs from product management, and the pivots — engineering manager, product manager, and AI-program leadership. Five moves, a judgment exercise, and a plan.

AI for Scrum Masters

Mon, 27 Jul 2026 00:00:00 GMT

An honest handbook, because this role is under more pressure than most: standups, sprint analytics, backlog grooming, burndown reports, and retro summaries are exactly what AI automates — and much of the pure ceremony-facilitation scrum-master job goes with them. But coaching a team, removing organizational impediments, resolving conflict, and driving actual delivery do not automate. What AI takes over, what stays valuable, and the honest pivots — delivery lead, technical program manager, engineering manager, and org-level agile coaching. Five moves, a judgment exercise, and a plan.

AI for Developer Relations

Mon, 27 Jul 2026 00:00:00 GMT

AI drafts your blog posts, scaffolds tutorials, writes the sample code, and answers routine developer questions — and, bigger, developers increasingly ask an AI instead of reading your content at all. But that same shift creates two new high-value jobs only DevRel can do: making your product the answer the AI gives (GEO, llms.txt, docs-as-agent-context) and building the authentic community and trust no model can fake. What AI automates versus what becomes valuable, and the pivots — product marketing, product management, docs/DX engineering, and AI-visibility (GEO) specialist. Five moves, a judgment exercise, and a plan.

AI for Founders

Mon, 27 Jul 2026 00:00:00 GMT

AI collapses the cost of building — one person or a tiny team can now ship what used to take a whole engineering org. That is real leverage, but it cuts both ways: when everyone can build, building stops being the moat. What AI makes cheap (code, MVPs, first-draft everything, research) versus what becomes scarce and decisive (taste on what to build, distribution, a real wedge, customer truth, and judgment on what NOT to build), how to use AI leverage without fooling yourself, and where to place your bet. Five moves, a judgment exercise, and a plan.

AI for Game Developers

Mon, 27 Jul 2026 00:00:00 GMT

AI generates concept art, textures, 3D assets, boilerplate gameplay code, dialogue drafts, and procedural content — and it is reshaping game production faster than almost any creative field, asset pipelines first. What AI automates (asset production, boilerplate, first-draft narrative, playtest analysis) versus what becomes valuable (game design and game feel, original creative vision, knowing what is actually fun, and building AI into the game itself), plus the pivots — technical game designer, AI-gameplay engineer, and AI-leveraged indie. Five moves, a judgment exercise, and a plan.