Skip to content
[ FIELD REFERENCE · AUTHORED DEFINITIONS ]

The forward deployed engineering field reference.

This reference defines the forward deployed engineering discipline as LockedIn Labs has authored it. Every term below is generated from the FDE Field Standard, the Pod Construct, and the Mission Hours policy, not written separately for this page, so a definition here cannot say something the underlying standard does not.

The discipline and the method

What a forward deployed engineer is, and the six phase method that structures the work.

Forward deployed engineering#
A forward-deployed engineer converts an ambiguous, consequential workflow into an adopted production system inside the customer's operating constraints, proves its behavior, transfers ownership, and returns reusable learning to the product without binding the solution to one model or interface.
Qualification is earned only from a coherent evidence chain: observed work under changing constraints, executable artifacts, independent review against published rubrics, a live defense, and supervised field evidence. Attendance, content completion, tool usage, and multiple-choice recall never satisfy the standard on their own.
Forward-Deployed Engineering Standard FDE-S/1.0, definition
The DEPLOY Method#
Diagnose the work. Establish the outcome contract. Prove the smallest wedge. Land inside the enterprise. Operate from evidence. Yield ownership and compound learning.
Diagnose asks: How does the work actually move, especially where the documented process breaks? Establish asks: What is the smallest safe wedge that can produce decision-grade evidence of workflow impact? Prove asks: Can the smallest wedge meet the mission contract under representative, edge, adversarial, and changing conditions? Land asks: Can the proven wedge live inside enterprise identity, data, control, and operating boundaries without losing its evidence? Operate asks: Can users adopt the system while the team detects, contains, and learns from behavior in production? Yield asks: Can the named owner operate, change, and recover the system—and what permitted learning should compound?
Forward-Deployed Engineering Standard FDE-S/1.0, DEPLOY method

Competency domains

The nine portable capability domains a forward deployed engineer is measured against.

FDE-01 · Field discovery and workflow reconstruction#
Reconstruct how work actually moves across people, systems, data, exceptions, incentives, and informal controls before proposing a solution.
Observes representative work and distinguishes stated process from actual process. Maps actors, systems of record, handoffs, wait states, exceptions, rework, and decision rights. Surfaces contradictory stakeholder accounts and resolves them with evidence rather than averaging them away. Separates symptoms from root constraints and records unknowns that can invalidate the mission.
Forward-Deployed Engineering Standard FDE-S/1.0, competency FDE-01
FDE-02 · Outcome framing and technical scoping#
Turn an ambiguous mission into a falsifiable outcome contract and sequence the smallest safe wedge that can reach production-shaped evidence.
Defines the user, decision, baseline, target condition, guardrails, and evidence window. Chooses agentic or deterministic methods from workflow characteristics rather than fashion. Negotiates scope, speed, quality, and risk tradeoffs explicitly with decision owners. Maintains an assumptions and kill-criteria register as evidence changes.
Forward-Deployed Engineering Standard FDE-S/1.0, competency FDE-02
FDE-03 · Enterprise architecture and integration#
Design a production-shaped system that respects identity, data lineage, legacy interfaces, transaction boundaries, operational ownership, and reversible change.
Traces data from authoritative source through transformation, model context, action, and retention. Designs around identity, least privilege, tenancy, rate limits, idempotency, failure recovery, and legacy constraints. Draws explicit deterministic, model, human, and external-system boundaries. Uses staged integration seams and rollback points rather than a big-bang replacement.
Forward-Deployed Engineering Standard FDE-S/1.0, competency FDE-03
FDE-04 · Provider-portable AI systems engineering#
Engineer context, instructions, tools, state, and control flow as testable contracts that can survive a change of model, provider, or runtime.
Selects a model and architecture from measured task needs, not a default vendor preference. Starts with the simplest adequate deterministic or single-agent design and adds coordination only when evidence requires it. Keeps provider-specific capabilities behind narrow adapters with normalized errors and traces. Runs an unchanged task harness across alternative configurations and records behavioral differences.
Forward-Deployed Engineering Standard FDE-S/1.0, competency FDE-04
FDE-05 · Evaluation, verification, and reliability#
Convert the mission contract into layered, repeatable evidence that detects behavior regressions across outputs, trajectories, system outcomes, and operating conditions.
Builds representative cases from real workflow patterns, including edge, adversarial, and abstention cases. Combines deterministic checks, model-based graders, human review, and production signals according to consequence. Inspects traces and failure clusters before changing prompts, models, tools, or architecture. Uses fixed thresholds, regression canaries, and explicit HOLD or SHIP decisions.
Forward-Deployed Engineering Standard FDE-S/1.0, competency FDE-05
FDE-06 · Security, governance, and responsible authority#
Bound what the system can know and do, connect controls to mission risk, and preserve evidence sufficient for authorization, investigation, and revocation.
Threat-models data exposure, prompt and tool injection, excessive agency, identity misuse, and supply-chain failure. Defines authority by action, resource, environment, duration, and escalation path. Maps each material risk to prevention, detection, response, evidence, and an accountable owner. Designs approval, abstention, kill-switch, retention, and audit behavior before production access expands.
Forward-Deployed Engineering Standard FDE-S/1.0, competency FDE-06
FDE-07 · Production operations and incident response#
Launch progressively, measure the end-to-end service, respond to degraded behavior, and improve reliability without hiding model uncertainty behind infrastructure health.
Defines service indicators for accepted outcomes, critical failures, latency, cost, review load, and adoption. Uses shadow, limited-authority, canary, and staged rollout modes with explicit promotion and rollback gates. Runs incident command from observable evidence and separates containment from root-cause remediation. Promotes field failures into evals, runbooks, controls, and ownership changes.
Forward-Deployed Engineering Standard FDE-S/1.0, competency FDE-07
FDE-08 · Client leadership and adoption#
Create shared clarity across users, executives, engineering, security, legal, and operations so that a technically sound system changes real work responsibly.
Adapts the same evidence into operator, engineering, risk, and executive decision formats without changing the facts. Runs milestone, risk, scope-change, and decision forums with named owners and dates. Designs training, support, workflow change, feedback, and adoption measurement into rollout. Escalates material disagreement and risk directly rather than manufacturing consensus.
Forward-Deployed Engineering Standard FDE-S/1.0, competency FDE-08
FDE-09 · Ownership transfer and field-to-product learning#
Make the client capable of operating and changing the system, then turn permitted field evidence into reusable, versioned product and delivery learning.
Names an internal owner early and requires that owner to operate, modify, and recover the system before exit. Transfers source, decisions, evals, controls, runbooks, known limits, and open risks in usable form. Separates client-specific evidence from reusable patterns and records permission and provenance before reuse. Writes field feedback as reproducible evidence tied to product, model, or method decisions.
Forward-Deployed Engineering Standard FDE-S/1.0, competency FDE-09

Evidence gates

The ten checkpoints an engineer's evidence must pass on the way to qualification.

EG-00 · Unassisted baseline work sample#
A time-bounded mission fragment establishes how the engineer discovers, scopes, architects, and verifies before instruction begins.
The submission is complete enough to locate strengths and gaps; no minimum score converts this diagnostic into qualification. Assistance, time, and tool conditions are recorded.
Forward-Deployed Engineering Standard FDE-S/1.0, evidence gate EG-00
EG-01 · Discovery and outcome contract#
The engineer resolves an ambiguous client brief through live discovery and produces a traceable workflow diagnosis and falsifiable wedge.
Every material claim traces to observation, source evidence, or a labeled assumption. The wedge can reach production-shaped evidence without requiring a full transformation. A reviewer can reject the mission from its explicit guardrails or kill criteria.
Forward-Deployed Engineering Standard FDE-S/1.0, evidence gate EG-01
EG-02 · Enterprise integration slice#
A production-shaped vertical slice crosses real-shaped identity, data, tool, and human-control boundaries with reversible failure behavior.
A clean evaluator can reproduce the slice without undocumented manual intervention. External writes are idempotent or compensated, permission-bounded, and observable. The reviewer injects at least one integration failure the engineer must diagnose and contain.
Forward-Deployed Engineering Standard FDE-S/1.0, evidence gate EG-02
EG-03 · Provider-portable system#
One behavioral contract operates across two materially different provider or runtime configurations and exposes non-equivalent behavior honestly.
The unchanged harness runs against both configurations. Provider-specific features are isolated and documented rather than hidden in shared logic. The engineer makes and defends a selection or non-equivalence decision from evidence.
Forward-Deployed Engineering Standard FDE-S/1.0, evidence gate EG-03
EG-04 · Evaluation and reliability control#
The engineer detects a seeded behavior regression, holds release, repairs the cause and coverage, and reruns invalidated gates.
Task, trajectory, outcome, safety, and operational checks are proportionate to mission risk. The seeded regression is found without disclosing the scoring key. No threshold or expected result is edited merely to turn a failure green.
Forward-Deployed Engineering Standard FDE-S/1.0, evidence gate EG-04
EG-05 · Security and authority review#
The engineer defends the threat model, data boundary, control evidence, residual risk, and authority envelope under adversarial questioning.
Every consequential action has an identity, bounded authority, observable record, and tested denial or recovery path. Human approval steps name the decision, evidence, timeout, fallback, and accountable person. The engineer narrows or stops the system when evidence does not support authorization.
Forward-Deployed Engineering Standard FDE-S/1.0, evidence gate EG-05
EG-06 · Production, adoption, and incident command#
The engineer runs a staged release, reads outcome and adoption signals, and leads containment through a changing incident scenario.
Promotion depends on accepted outcomes and critical-failure bounds, not infrastructure health alone. The engineer separates containment from remediation and communicates uncertainty without delay. The incident changes at least one test, control, runbook, or ownership decision.
Forward-Deployed Engineering Standard FDE-S/1.0, evidence gate EG-06
EG-07 · Supervised field mission#
In a real client boundary, the engineer owns one bounded mission from discovery through a separate candidate-owned production release and observed operation while a supervisor retains escalation authority.
A separate candidate-owned release reached a real production environment and was observed in real use; simulation or production-equivalent rehearsal cannot satisfy this gate. The work uses the client's actual constraints and a named decision owner. The engineer changes course when field evidence contradicts the plan. A named owner can operate and recover the released system without hidden candidate intervention. Client-specific material remains inside the agreed rights and retention boundary.
Forward-Deployed Engineering Standard FDE-S/1.0, evidence gate EG-07
EG-08 · Ownership transfer and field learning#
A named owner operates, changes, and recovers the system while the engineer defends known limits and submits permission-bounded field learning.
The named owner completes the run without hidden FDE intervention. Open risk, debt, and deferred decisions carry owners and dates. Reusable learning excludes client-confidential evidence unless separate permission exists.
Forward-Deployed Engineering Standard FDE-S/1.0, evidence gate EG-08
EG-09 · Integrated qualification defense#
The engineer defends the complete evidence chain, handles a live scope change, diagnoses an injected failure, and explains the mission to technical, risk, operator, and executive audiences.
Every competency is supported by independently reviewed evidence rather than completion metadata. The engineer can explain tradeoffs, limitations, and contradictory evidence without scripts. Critical-failure criteria are conjunctive: strength in one domain cannot average away unsafe failure in another.
Forward-Deployed Engineering Standard FDE-S/1.0, evidence gate EG-09

Qualification stages

The five cumulative claims the evidence gates support, from a recorded baseline to a defended qualification.

QS-00 · Baseline established#
The engineer has a recorded starting profile and an evidence-based development plan.
Diagnostic only. It establishes no readiness, field capability, or credential eligibility.
Forward-Deployed Engineering Standard FDE-S/1.0, qualification stage QS-00
QS-01 · Mission-ready in simulation#
The engineer can diagnose, scope, and integrate a bounded FDE mission in a controlled environment.
Simulation evidence does not establish production operation or performance inside a client environment.
Forward-Deployed Engineering Standard FDE-S/1.0, qualification stage QS-01
QS-02 · Production-ready in rehearsal#
The engineer has demonstrated evaluation, authority, rollout, adoption, and incident judgment under controlled pressure.
A rehearsal is not a production outcome and does not replace supervised client work.
Forward-Deployed Engineering Standard FDE-S/1.0, qualification stage QS-02
QS-03 · Field evidence complete#
The engineer has owned a bounded mission in a real client boundary and transferred operation to a named owner under supervision.
One supervised mission does not predict performance in every domain, industry, scale, or employer context.
Forward-Deployed Engineering Standard FDE-S/1.0, qualification stage QS-03
QS-04 · FDE qualification defended#
The complete, current evidence chain has passed an independent integrated defense across all nine domains.
Qualification is an evidence-backed program judgment, not a job guarantee, employer authorization, security clearance, or proof of outcomes beyond the reviewed missions.
Forward-Deployed Engineering Standard FDE-S/1.0, qualification stage QS-04

Pod seats

The named roles inside and around a forward deployed engineering pod.

FDE Lead#
The outcome, the cadence, and the stakeholders. Carries the engagement management function, so there is no separate project manager layer.
Level principal. Posture onsite. Placement in-pod. A billable seat. Comparable market titles include Engagement lead, Delivery lead, Consulting lead. Every pod is anchored by an onshore lead. Never an anchorless pod.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, seat fde-lead
AI Solution Architect#
System design, estate integration, and model strategy. One architect can span two pods.
Level principal. Posture hybrid. Placement in-pod. A billable seat. Comparable market titles include GenAI architect, AI/ML solution architect. One person in this seat may span up to 2 pods.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, seat fde-architect
Senior FDE#
Agent design, typed tooling, and the hardest workflows. Mentors the FDE I seats in the pod.
Level fde-ii. Posture hybrid. Placement in-pod. A billable seat. Comparable market titles include Senior GenAI engineer, Agentic AI engineer.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, seat senior-fde
FDE#
LLM based workflows, copilots, and automation built inside the client estate.
Level fde-i. Posture hybrid. Placement in-pod. A billable seat. Comparable market titles include Forward Deployed Engineer, GenAI engineer, AI engineer.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, seat fde
Associate FDE#
Bounded build work under a lead. An early career engineer trained and evidenced through this platform, always deployed paired with a lead and never alone.
Level associate. Posture onsite. Placement in-pod. A billable seat. Comparable market titles include Junior GenAI engineer, Associate AI engineer. The visible front of the training pipeline and a lower cost onshore seat.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, seat associate-fde
FDE + AI Evaluation#
Evaluation sets, protected data gate evidence, and shadow mode readouts. The seat that makes governed AI provable.
Level fde-i. Posture fractional. Placement in-pod. A billable seat. Comparable market titles include AI quality engineer, QA, modernized. Fractional or dedicated. A standard pod flexes to four seats by making this seat fractional.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, seat fde-evaluation
Data and Integration bench#
Pipelines and connectors so pod engineers stay unblocked. Shared across pods, off the order.
Level bench. Posture remote. Placement behind-pods. Not a billable seat. Comparable market titles include Data engineer, Integration engineer.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, seat data-integration-bench
Security and Compliance SME#
CISO led review and audit evidence when the work demands it.
Level on-call. Posture on-call. Placement behind-pods. Not a billable seat. Comparable market titles include Security engineer, GRC specialist.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, seat security-compliance-sme
Agile Delivery Manager#
The scrum function for one pod, only when an order specifies it. By default the lead carries team ceremonies.
Level coordination. Posture hybrid. Placement coordination. A billable seat. Comparable market titles include Scrum master.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, seat agile-delivery-manager
Agile Program Manager#
Cross pod orchestration for three to five pods: dependencies, program cadence, and consolidated reporting.
Level coordination. Posture hybrid. Placement coordination. A billable seat. Comparable market titles include Release train engineer.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, seat agile-program-manager
Engagement Executive#
Ownership of the program relationship, onsite for setup and critical phases. Included across all pods at no per pod cost.
Level program. Posture hybrid. Placement program-layer. Not a billable seat. Comparable market titles include Executive sponsor.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, seat engagement-executive
Chief FDE Architect#
Architecture and method consistency across every LockedIn Labs pod in the program. Included across all pods at no per pod cost.
Level program. Posture hybrid. Placement program-layer. Not a billable seat. Comparable market titles include Chief architect.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, seat chief-fde-architect
Product Owner#
Product ownership belongs with the business. LockedIn Labs slots under the client's product structure and never duplicates it.
Level client. Posture onsite. Placement client-side. Not a billable seat. Comparable market titles include Product owner, Product manager.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, seat product-owner

Ladder rungs

The five rungs an engineer climbs, each earned on recorded qualification stages rather than tenure.

Associate FDE#
A recorded baseline and an evidence based development plan. Deploys only paired with a lead.
Diagnostic standing. It establishes no independent field capability. Requires the following qualification stages: Baseline established.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, ladder rung associate-fde
FDE I#
Can diagnose, scope, integrate, evaluate, and stage a bounded mission under controlled pressure.
Rehearsal evidence is not a production outcome inside a client environment. Requires the following qualification stages: Baseline established, Mission-ready in simulation, Production-ready in rehearsal.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, ladder rung fde-i
FDE II (Senior)#
Has owned a bounded mission in a real client boundary and transferred operation to a named owner under supervision.
One supervised mission does not predict performance in every domain, scale, or estate. Requires the following qualification stages: Baseline established, Mission-ready in simulation, Production-ready in rehearsal, Field evidence complete.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, ladder rung fde-ii
FDE Lead#
The complete evidence chain has passed an independent integrated defense and the engineer anchors a pod onshore.
A defended qualification is a program judgment, not a client acceptance or an employer authorization. Requires the following qualification stages: Baseline established, Mission-ready in simulation, Production-ready in rehearsal, Field evidence complete, FDE qualification defended.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, ladder rung fde-lead
FDE Architect#
The complete evidence chain has passed an independent integrated defense with architecture and integration depth across estates.
A defended qualification is a program judgment, not a client acceptance or an employer authorization. Requires the following qualification stages: Baseline established, Mission-ready in simulation, Production-ready in rehearsal, Field evidence complete, FDE qualification defended.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, ladder rung fde-architect

Pod shapes

The authored team sizes a pod can take, from one engineer to a multi pod program.

Solo FDE#
A single senior FDE embedded in an existing team: requirements, build, integration, and stakeholder management in one person. The fastest way to add AI depth to an existing squad or own a scoped workflow end to end.
One engineer, whole loop. Composed of 1 × Senior FDE. Sized from 1 to 1 people.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, pod shape solo-fde
Embedded Pair#
Assessment, the estate map, and the first working workflow. Two people, two to four weeks, and a credible plan grounded in working software.
FDE plus architect. Composed of 1 × AI Solution Architect, 1 × Senior FDE. Sized from 2 to 2 people.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, pod shape embedded-pair
Build Pod#
The standard unit for delivery programs: lead, architect, and builders with the full method inside. Complete, anchored, or blended with the client's seats.
Five seats, anchored. Composed of 1 × FDE Lead, 1 × AI Solution Architect, 1 × Senior FDE, 1 × FDE, 1 × FDE + AI Evaluation. Sized from 4 to 9 people.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, pod shape build-pod
Pod Program#
Spanning architects, shared evidence gates, weekly readouts, and builder waves through the training pipeline. Program leadership included.
Many pods, one layer. Composed of 1 × FDE Lead, 1 × AI Solution Architect, 1 × Senior FDE, 1 × FDE, 1 × FDE + AI Evaluation. Sized from 4 to 9 people.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, pod shape pod-program

Pod configurations

The three ways seats split between LockedIn Labs and a client or partner.

Configuration A · Complete pod#
All seats LockedIn Labs. Full method, single accountability, fastest to production quality.
LockedIn Labs holds FDE Lead, AI Solution Architect, Senior FDE, FDE, FDE + AI Evaluation. Flex seats: none. Method: runs end to end.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, configuration A
Configuration B · Anchored pod#
LockedIn Labs holds lead, architect, and one to two builders. Flex seats stay the client's or a partner's; the pod still runs the method end to end.
LockedIn Labs holds FDE Lead, AI Solution Architect, Senior FDE. Flex seats: FDE, FDE + AI Evaluation, Associate FDE. Method: runs end to end.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, configuration B
Configuration C · Certified seats#
Individual evidenced FDEs into pods the client runs. Fastest fill for an order; an architect stays on call across seats.
LockedIn Labs holds Senior FDE, FDE. Flex seats: FDE Lead, AI Solution Architect, FDE + AI Evaluation, Associate FDE. Method: architect on call.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, configuration C

Operating frame

The boundary between the client's delivery process and the method LockedIn Labs runs inside the pod.

Operating frame#
Outside, your process. Inside, our method.
Your operating frame covers Product owner, Scrum cadence, Backlog, Ceremonies, Program reporting, owned by the client. SAFe compatible. LockedIn Labs slots under the client's product structure and never duplicates it. Inside the pod sequences Spec and PRD, Agentic build, Protected data gate, Evaluation sets, Human gate, Shadow mode, Production, owned by LockedIn Labs.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, operating frame
Spec and PRD#
Context engineered before any build starts.
Kind: stage. Sequenced inside the operating frame's inner method.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, operating frame stage spec
Agentic build#
Typed tools only. Requirements, design, code, and tests generated as auditable artifacts.
Kind: stage. Sequenced inside the operating frame's inner method.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, operating frame stage agentic-build
Protected data gate#
Protected data never reaches a model without policy clearance.
Kind: gate. Sequenced inside the operating frame's inner method.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, operating frame stage phi-gate
Evaluation sets#
Measured behavior against versioned cases and thresholds.
Kind: stage. Sequenced inside the operating frame's inner method.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, operating frame stage evaluation-sets
Human gate#
A person signs every consequential decision.
Kind: gate. Sequenced inside the operating frame's inner method.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, operating frame stage human-gate
Shadow mode#
Live data, no authority, until the readout holds.
Kind: stage. Sequenced inside the operating frame's inner method.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, operating frame stage shadow-mode
Production#
Signed live by the accountable officer.
Kind: release. Sequenced inside the operating frame's inner method.
Forward Deployed Engineering Pod Construct FDE-POD/1.0, operating frame stage production

Mission hours workstreams

The six categories of supervised work counted toward field entrustment.

MH-01 · Discover and quantify the opportunity#
Can the engineer observe the workflow, locate the real constraint, quantify consequence, and recommend no build when AI is the wrong intervention?
Workflow, exception, wait-state, and decision-rights map Sourced baseline, value hypothesis, and no-build alternative Stakeholder interview record and bounded mission charter Minimum authored hours: 20.
Mission Hours and Field Entrustment FDE-MH/0.1, workstream MH-01
MH-02 · Scope architecture, data, delivery, and controls#
Can the engineer turn ambiguity into a reversible system boundary with explicit authority, data, provider, cost, and operating choices?
Outcome contract, authority envelope, and system context Architecture decisions, provider adapter, and threat model Acceptance, kill, escalation, and rollback criteria Minimum authored hours: 20.
Mission Hours and Field Entrustment FDE-MH/0.1, workstream MH-02
MH-03 · Build and integrate the production-shaped wedge#
Can the engineer implement the smallest defensible slice across real interfaces without hiding brittle assumptions behind a polished demo?
Executable artifact with typed interfaces and bounded tools Identity, data, observability, and portability evidence Change record tying implementation decisions to the mission contract Minimum authored hours: 60.
Mission Hours and Field Entrustment FDE-MH/0.1, workstream MH-03
MH-04 · Evaluate quality, safety, security, and governance#
Can the engineer prove behavior against exact candidates, adversarial cases, human-review cost, authority limits, and critical failure conditions?
Versioned evaluation set with held-out and adversarial cases Security, privacy, and excessive-agency review Go, hold, repair, and rerun decision trail Minimum authored hours: 30.
Mission Hours and Field Entrustment FDE-MH/0.1, workstream MH-04
MH-05 · Deploy, observe, operate, and recover#
Can the engineer own an exact release into a real production environment, read evidence from real users or operators, contain an incident, recover safely, and prove that the repaired system remains inside bounds?
Candidate-owned repository, commit, immutable artifact, production environment, staged launch, and observability record Real-user or real-operator observation with accountable ownership, SLO, outcome, guardrail, cost, and adoption evidence Incident or game-day decision chronology Containment, recovery, rerun, and post-incident evidence Minimum authored hours: 30.
Mission Hours and Field Entrustment FDE-MH/0.1, workstream MH-05
MH-06 · Drive adoption, executive decision, and owner transfer#
Can the engineer earn a bounded business decision, measure adoption and value, handle objections, and leave a named owner able to operate the system?
Adoption-adjusted business case and executive decision memo Objection, scope-change, procurement, and support record Named-owner teach-back, owner-executed operation and recovery, runbook, transfer acceptance, and 30-day backlog Minimum authored hours: 20.
Mission Hours and Field Entrustment FDE-MH/0.1, workstream MH-06
02[ SEE IT IN THE PRODUCT ]

Where these terms are demonstrated.

A definition is one thing. Watching the same term operate inside the product is another. Each link below opens the surface where the term above is put to work.

03[ SOURCE RECORDS ]

Exactly what this reference reflects.

This page is generated, not hand-written. Every term traces to one of the following authored records, at the exact version listed.

Forward-Deployed Engineering Standard
FDE-S/1.0 · version 1.0.1
Forward Deployed Engineering Pod Construct
FDE-POD/1.0 · version 1.0.0
Mission Hours and Field Entrustment
FDE-MH/0.1 · version 0.1.2

A definition here is not a claim about any person, engagement, or outcome. It records what the standard specifies, not who has met it, what any engagement produced, or how any individual performed.