Senior AO Engineer

P. Moe Talent Collective; United Kingdom; Remote; GBP 90,000 - 140,000; Full time

Job summary

P. Moe Talent Collective is seeking a Senior Software Engineer (AI-Native / Python & TypeScript) for a UK-based LegalTech platform. You’ll help design and implement agent-driven production software in small pods, with strong emphasis on correctness and reliability.

Location is predominantly remote with occasional London meetings; immediate start. You’ll work with a cross-functional team including a Product/Spec Lead, an Embedded Domain Expert, and Verification Engineer, shaping the new

Qualifications

• Proven experience with agent-first development in production.
• Experience in planning, decomposing and directing agent-driven work.
• Strong Python and TypeScript engineering skills at production scale.
• Excellent written and spoken English with precise technical communication.

Responsibilities

• Plan requirements, dependencies, and acceptance criteria; decompose work into agent-executable tasks.
• Orchestrate coding agents, write precise instructions, manage context and parallel tasks.
• Review agent-generated code against specs; support verification and security reviews.
• Deliver small, safe, reviewable changes with high first-pass acceptance.
• Collaborate across disciplines and communicate plans, decisions, and risks.

Skills

Python

TypeScript

Agent-first Dev

Software architecture

Code verification

English communication

Collaboration

Tools

Claude Code

Codex

Copilot

Job description

Senior Software Engineer (AI-Native / Python & TypeScript)

My client builds a specialist LegalTech platform for complex legal and private funds workflows and is moving to an agent-first engineering model to deliver production software considerably faster while improving quality, correctness, and reliability.

These are not roles where AI merely assists: engineers spend ~90% of effort specifying, directing, and verifying coding agents and <10% writing code, yet remain accountable for the technical approach and code quality.

You will work in small, specialist delivery pods (5–6 people) supported by a Product/Spec Lead, an Embedded Domain Expert, and a Verification Engineer, underpinned by centralised architecture, security, agent harness, and platform/SRE expertise. You will own how approved requirements are implemented and champion customers' needs. As one of the first hires in an immediate hiring drive, you will help establish the model and prove that small teams can substantially out-deliver conventional ones in volume, correctness, and quality.

My client is tooling agnostic (Claude Code, Codex, Cursor, Copilot, Windsurf, Devin, etc.); what matters is how effectively you use them.

KEY RESPONSIBILITIES

• Planning: Clarify requirements, constraints, dependencies, and acceptance criteria (consulting on problem framing, discovery, example mapping, and spec writing); surface ambiguities and risks before build; choose the approach and decompose work into agent-executable tasks with clear instructions and boundaries; hold accountability for technical soundness at plan approval and make significant decisions within architectural, security, and platform standards.
• Agent Orchestration: Use coding agents as the principal implementation method; hold accountability for task decomposition and agent runs; write precise instructions; manage context and coordinate multiple agents concurrently; catch incorrect assumptions and defects, redirect agents, and resolve conflicts between tasks without reverting unnecessarily to manual coding; help build repeatable practices.
• Engineering Judgement: Become productive quickly in complex existing codebases; assess correctness, maintainability, performance, security, and reliability; spot architectural problems, flawed assumptions, and breaches of engineering constraints or agent-harness rules; personally read, debug, write, and refactor the hard 10% of difficult production code by hand where agents fall short.
• Review & Verification: Hold accountability for Tier 2–3 code reviews, reviewing agent-generated code against specifications and standards alongside Architecture and Security; consult on acceptance test design and work closely with Verification Engineers (who own test strategy, test oracles, and acceptance evidence); use automated verification, investigate failures, and direct fixes; address security and specialist review findings.
• Delivery: Ship small, safe, reviewable changes with high first-pass acceptance and minimal rework; consult on deploy, rollout, and rollback; clear obstacles fast; demonstrate meaningful gains in delivery speed and productivity without compromising quality.
• Collaboration: Communicate plans, decisions, and risks clearly; challenge constructively; consult on threat models per epic, non-functional constraints, and the central agent harness, context, and skills; share feedback on tooling, workflows, and verification; adapt as processes and team structures evolve; build a culture of excellence, accountability, and continuous improvement.

ESSENTIAL REQUIREMENTS

• Agent-First Development: An established professional practice of using coding agents as the primary way to build production software, working substantially towards the 90/10 model. Occasional AI assistance, or building AI-powered products, does not qualify.
• Production Verification: Experience reviewing, testing, and validating agent-generated production code (strong conventional automated-verification experience considered if the agent-first requirement is met). Owning a verification platform is not required.
• Python and TypeScript: Professional capability in both (need not be equal). Strong backend engineers welcome; full stack not required.
• Complex Production Software: Track record of understanding and extending mature codebases, their architecture, and technical debt. Predominantly greenfield experience is insufficient.
• Ownership and Judgement: Personally made significant technical decisions and owned the outcome (people management or platform ownership not required).
• Decomposition and Comprehension: Break complex problems into well-scoped tasks with the right context; understand unfamiliar code quickly; solve hard problems manually when needed.
• Proven Productivity Gains: Specific, credible examples of improved delivery through agent-first development (formal metrics not essential).
• Communication: Exceptional written and spoken English—precise, unambiguous technical instruction is critical when directing agents.
• Working Style: Comfortable at pace in an evolving environment; pragmatic, adaptable, and collaborative across disciplines.

HIGHLY DESIRABLE

• Concurrent multi-agent orchestration (exceptional single-agent engineers considered if they can explain how they would manage parallel work and context switching).
• Advanced context management and parallel workflows.
• Implementing agent-first practices, guardrails, or automated verification.
• Modernising/refactoring complex systems.
• Strong technical writing; background or interest in language, linguistics, reading, or etymology.
• Exposure to agent harnesses, MCP, verification infrastructure, security tooling, or SRE is useful but not core.
• Any software sector and any company size (start-up to enterprise)—seniority is judged on depth, judgement, and achievement, not years or title.

WHAT TO EXPECT IN DISCUSSIONS

When speaking with us, please be ready to discuss specific examples of:

• Agent-first development and your agent/manual balance.
• Assessing requirements and decomposing problems.
• Directing agents, managing context, and parallel workflows.
• Significant technical decisions you made.
• Reviewing agent code and establishing correctness through automated verification.
• Hard problems you solved when agents fell short.
• Productivity gains achieved.
• Collaborating with specialists through changing practices.
• Communicating complex technical information precisely.

Reports to: Director of Engineering

Location: Predominantly remote, occasional London meetings (London-commutable preferred; other locations considered)

Availability: Immediate need to hire; short-notice availability advantageous