Senior AI Architect

Title: Senior AI Architect

Employment type: Full-time

Domains: Banking & Financial Services, Insurance, Travel, Healthcare, Hi-Tech, Retail

What you'll own

This role is measured on outcomes delivered, not tickets closed or hours billed.

• Own a client outcome end to end. The business result the client is paying for, from ambiguous first conversation to production system.
• Command a pod. Associate FDEs and AI agents are your team from day one. You set technical direction, review their work, and are accountable for what the pod ships.
• Architect and ship into production. Inside the client's actual security model, data residency rules, and legacy stack.
• Operate as the engagement's de facto CTO. Translate an ambiguous client problem into scope, and scope into shipped software, without waiting for someone else to define the spec.
• Leave the client's team stronger. Transfer capability and documentation. A good engagement reduces the client's dependency on us, not increases it.
• Say the hard thing. To the client, when the scope is wrong or the timeline is fiction.

Who thrives in this role

• 6–10 years shipping production software, with meaningful time in customer-facing, embedded, or consulting-adjacent delivery.
• Has already put an agentic system into production inside a regulated or otherwise constrained environment — not a demo, not a hackathon.
• Has led a small team, formally or informally, and is comfortable being one of a handful of senior people that associate engineers look to for direction.
• Uses agentic coding tools as a default working method, not a novelty.
• Has sat across the table from a client's engineering, security, and compliance teams simultaneously, and gotten to yes.
• Wants to be evaluated on outcomes delivered, not hours logged.

What we evaluate, and where

We're evaluating

What "good" looks like

Where it's tested

Production engineering depth

Has shipped and operated systems at scale; reasons about failure modes, not just the happy path

Technical work sample

Applied AI / agentic systems

Has built with LLMs and agents beyond prompting — RAG, tool use, evals, orchestration

Technical work sample + technical interview

AI-native working style

Uses agentic tools as a default method of building, evaluated live

Live pairing session, your own tools

Regulated-industry delivery

Has navigated SSO, data residency, model risk, or equivalent real-world constraints

Case interview

Client-facing judgment

Separates signal from noise in an ambiguous, high-stakes client ask

Case interview

Leadership and teaching ability

Explains the same technical decision differently to a junior engineer and to a client CTO

Pod-fit conversation with current FDEs

AI fluency is a baseline requirement here, tested directly.

What this role is not

• Not staff augmentation. You own an outcome, not a backlog assigned by a client PM.
• Not advisory. You ship production code. If the engagement doesn't produce shipped software, it doesn't count.
• Not tied to one vendor. You'll work across Anthropic, OpenAI, Google, SpaceXAI, and NVIDIA stacks depending on what the client's problem requires — not a single-platform mandate.
• Not remote-first. This role lives at the client site during active engagements.