AI roles hub

Eight AI roles. One precise hire.

Choose the ownership model your roadmap actually needs. Every Devlyn role has public pricing, AI-native vetting, a 48-hour shortlist, and a two-week paid trial in your codebase.

Book a 30-minute role scopeNeed a guide?
Diagnostic

Match by bottleneck, not title.

Role map

Pick the role by first proof.

Every card names the symptom, ownership, first 14-day proof, and starting price.

FDE$2,500/mo+

AI feature stuck between demo and production?

Forward-Deployed AI Engineer

Best when: AI demos stall when no one owns the messy middle between customer reality, product design, model behavior, integrations, security, and release.

First proof: Workflow map, feature-flagged slice, eval baseline.

View role
APP$2,500/mo+

Need product UI on top of LLMs?

AI Application Engineer

Best when: The model may work, but users do not buy notebooks. They need reliable product flows, clear states, fast interactions, fallback behavior, and features that fit the existing app.

First proof: Working feature slice, streaming UI, telemetry.

View role
LLM$2,500/mo+

Model quality, routing, evals, or fine-tuning unclear?

LLM Engineer

Best when: Teams ship prompt changes by feel, switch models without baselines, and cannot explain whether quality improved, cost rose, or regressions appeared.

First proof: Eval dataset, scorecard, cost/latency report.

View role
RAG$2,500/mo+

Model must answer from your private data?

RAG & Context Engineer

Best when: The assistant sounds confident, but it cannot prove the answer, misses relevant context, retrieves stale documents, leaks permissioned data, or hallucinates over weak evidence.

First proof: Retrieval baseline, citation design, relevance eval set.

View role
AGT$2,500/mo+

Need tool-using agents with approvals and retries?

Agentic Workflow Engineer

Best when: Agent demos look impressive until they hit real permissions, unreliable tools, partial failures, retries, approval rules, rate limits, and unclear accountability.

First proof: Workflow graph, first agent slice, trace logs.

View role
PLT$2,500/mo+

Every team is building separate AI wrappers?

AI Platform Engineer

Best when: Every team builds its own prompt wrappers, model access patterns, eval spreadsheets, vector stores, traces, and cost tracking. AI delivery becomes fragmented before it scales.

First proof: Gateway/platform slice, trace convention, cost baseline.

View role
SEC$2,500/mo+

AI security risk blocking release?

AI Security Engineer

Best when: Traditional application security does not fully cover prompt injection, indirect prompt attacks, tool misuse, sensitive context leakage, unsafe model outputs, and AI-specific logging/monitoring gaps.

First proof: Threat model, attack scenarios, security backlog.

View role
DSC$2,500/mo+

Need decisions, forecasts, experiments, or causal analysis?

Data Scientist

Best when: Teams collect data but still make roadmap, growth, risk, and product decisions by opinion because metrics are messy, experiments are weak, and models are not tied to business decisions.

First proof: Data quality audit, metric tree, baseline analysis.

View role

Adjacent roles

Close names can hide different work.

Use this matrix to avoid hiring the role that sounds right but proves the wrong thing.

CompareChoose the first whenChoose the second when
FDE vs AI ApplicationThe problem is ambiguous across customer, product, and release.The product spec is clear and the AI app layer needs implementation.
LLM vs RAGModel behavior, routing, evals, or fine-tuning is unclear.The answer must be grounded in private or permissioned data.
Agentic vs PlatformOne workflow needs tools, approvals, retries, and traces.Many teams need shared AI infrastructure and governance.
Security vs PlatformAI-specific attack surface is blocking release.Security controls need to be enforced through shared platform services.
Data Scientist vs LLMThe question is business measurement, forecasting, or experiments.The question is model behavior, evals, routing, or cost.

Pricing preview

The ladder is public before the call.

Junior
$2,500/mo

Supervised delivery for clear implementation work.

Mid
$3,500/mo

Independent feature ownership for production AI work.

Senior
$4,500/mo

High-judgment ownership for ambiguous or risky AI delivery.

Hiring process

Role scope, shortlist, trial, then continue.

Day 0

30-minute role scope

Map the AI workflow, current stack, first deliverable, security boundaries, seniority, and the role that should own the work.

Hour 48

2-3 vetted engineers

Receive a short list with matching rationale. The goal is fewer names with stronger fit, not resume volume.

Week 1-2

Paid trial in your codebase

The selected engineer works inside your repo, rituals, issue tracker, and review process so fit is judged by real work.

After trial

Continue, replace, pause, or scale

Continue month-to-month, request a free replacement, pause without a long lock-in, or add adjacent roles.

FAQ

Role-selection questions.

Which AI role should we hire first?

Start with the bottleneck: ambiguous rollout, app UX, model quality, retrieval, agents, platform, security, or decision science.

Can Devlyn help us choose?

Yes. The 30-minute role scope exists to map the workflow and confirm whether one of the eight roles fits.

Do all roles use the same pricing ladder?

Yes. Junior starts at $2,500/mo, mid at $3,500/mo, and senior at $4,500/mo.

Can we hire multiple roles?

Yes. Many teams start with one role, then add an adjacent specialist once the first trial proves fit.

What if we choose the wrong role?

Devlyn should catch that during scoping. If fit is wrong during the engagement, free replacement support is available.

Are these full-time engineers?

Yes. Devlyn is built around dedicated ownership, not fragmented hourly freelancing.