AI Development Lead

Evolutions; Singapore; On-site; SGD 120,000 - 180,000; Full time

Job summary

EvolutionsHQ is seeking a Product Owner/Engineer with hands-on experience in agentic tooling to enable fast, reliable AI-driven development. You will bridge core tooling and product delivery across three layers: tooling infrastructure, owner enablement, and feature rollouts, in a regulated environment.

You will work with Elixir/Phoenix-style stacks and developer tooling, focusing on robust guardrails, documentation, and scalable workflows that empower product owners and engineers alike.

Qualifications

• Spec-driven development guides work from analysis to delivery.
• Provide explicit context, conventions, and constraints so agents fit the codebase.
• Verify outputs with tests, contracts, and targeted reviews to catch issues.

Responsibilities

• Build and operate a multi-host agent environment with provisioning and session persistence.
• Design MCP servers, agent skills, and project context for effective tool use.
• Define guardrails, tests, and CI gates for AI-produced changes and auditability.
• Support feature rollouts with monitoring and rapid fixes in production.

Skills

Spec-driven development

Context engineering

Verification with tests

Sustainable velocity

Strong software engineering

Security awareness

Tools

Linux

Shell

Git

Docker

GitHub Actions

AWS

Job description

Evo is building the next generation of investment management infrastructure. Our mission is to leverage AI-native technology and operating experience for institutions navigating a changing investment industry. Connect the information, people, and workflows firms need to launch, operate, and grow their businesses- giving them a stronger foundation for managing complexity and serving investors.

We develop targeted solutions and work directly with organizations across investment management, fund services, financial infrastructure and digital markets. We build in the traditional regulated finance and emerging digital infrastructure spaces, with the operating experience to understand both. From fund operations and accounting to reporting and oversight, we’re building technology and services that bring fragmented processes together, reduce manual work, and make information easier to trust. Practical expertise shapes everything we build, with professional judgment and accountability at the heart of our approach.

About Evo

Evo is building the next generation of investment management infrastructure. Our mission is to leverage AI-native technology and operating experience for institutions navigating a changing investment industry. Connect the information, people, and workflows firms need to launch, operate, and grow their businesses- giving them a stronger foundation for managing complexity and serving investors.

We develop targeted solutions and work directly with organizations across investment management, fund services, financial infrastructure and digital markets. We build in the traditional regulated finance and emerging digital infrastructure spaces, with the operating experience to understand both. From fund operations and accounting to reporting and oversight, we’re building technology and services that bring fragmented processes together, reduce manual work, and make information easier to trust. Practical expertise shapes everything we build, with professional judgment and accountability at the heart of our approach.

About the role

Evo is built AI-native. Our product owners work directly in Claude Code/Codex terminals and ship features alongside our engineers. This role is what makes that model work.

You’ll sit between core tooling and product delivery, working across three layers:

• Core tooling and infrastructure. You’ll work directly with our technical founder on the agent platform itself: the fleet of hosts and terminals, parallel worktrees, MCP servers, skills, CI gates, and the guardrails that keep agent output safe to merge.
• Product owner enablement. You’ll make product owners effective in the terminal. That means onboarding them, writing playbooks and spec templates, unblocking them in real time, and turning friction you see repeated into tools.
• Product feature rollouts. You’ll help carry features from spec to production, reviewing and hardening agent-built work and then testing, releasing, and monitoring it so it’s fit for a regulated platform.

This isn’t a research role, and it isn’t a pure platform role. It’s for someone who loves building the tools and seeing other people get faster because of them.

Our stack

• Product: Elixir, Phoenix / LiveView, Ash Framework, PostgreSQL
• Infrastructure: AWS, GitHub Actions, Linux hosts, Docker, tmux, git worktrees
• Agentic development: Claude Code and Codex running parallel multi-agent workflows, spec-driven development with OpenSpec, and internal MCP servers that expose our tools and data to agents

What you’ll do

Tooling and infrastructure

• Build and operate our multi-host agent environment: provisioning, session persistence, worktree management, secrets, sandboxing, and usage/cost visibility.
• Design, build, and maintain MCP servers, agent skills, and project context so agents understand our codebase, conventions, and domain.
• Own the guardrails for agent-authored code: test and CI gates, evals, review automation, and a traceable audit trail for AI-authored changes.
• Evaluate new models, agents, and tools as they ship. Adopt what works, drop what doesn’t, and write down why.

Product owner enablement

• Onboard product owners into Claude Code/Codex and our spec-driven workflow, and level them up over time.
• Build the paved road: OpenSpec templates, reusable skills, runbooks, and examples that let non-engineers produce work engineers can merge.
• Act as first responder when a product owner is stuck, whether the cause is a confused agent, a broken environment, or a spec that needs sharpening.
• Spot patterns in where people struggle and fix them at the tooling level, not one person at a time.

Feature rollouts

• Review and harden agent-built features before they reach production, with extra care for the ledger, data integrity, and compliance-relevant paths.
• Write and improve tests, migrations, and deploy steps so features ship safely and can be rolled back.
• Support releases end to end: staging, feature flags, monitoring, and follow-up fixes.

What we’re looking for

• Deep, hands-on agentic development experience. You use Claude Code, Codex, or similar tools daily on real production work, including running several agents in parallel. You can show:
• Spec-driven development: you break work into specs, tasks, and acceptance criteria before handing it to an agent.
• Context engineering: you give agents the right context, conventions, and constraints so their output fits the codebase.
• Verification over trust: you catch plausible-but-wrong code with tests, contracts, and targeted review, and you know the common failure modes (hallucinated APIs, silent behavior changes, over-engineering, security regressions).
• Sustainable velocity: you ship faster without piling up unreviewed code or debt nobody understands.
• Strong software engineering fundamentals. Production experience in at least one of Elixir, Python, or TypeScript, and willingness to work in Elixir day to day.
• Developer infrastructure skills: Linux, shell, git (including worktrees), Docker, CI/CD with GitHub Actions, and AWS.
• Tool-building experience. You’ve built internal tools, CLIs, integrations, or MCP servers that other people actually use.
• A teaching instinct. You can explain agentic workflows to a product owner as clearly as to an engineer, and you’re patient when someone is learning.
• Security judgment for AI tooling: you’re aware of prompt injection, secret leakage, over-broad agent permissions, and what data should never leave our environment.
• Comfort working in a regulated or high-stakes domain, where "mostly right" isn’t good enough.

Nice to have

• Elixir, Phoenix/LiveView, or Ash Framework experience.
• PostgreSQL depth: schema design, migrations, and query performance.
• Experience in developer experience, platform engineering, or a developer-tools company.
• Experience writing evals for LLM-backed features, or shipping LLM features to production with cost, latency, and accuracy controls.
• A background in fintech, financial services, or compliance (SOC 2, KYC/AML, audit requirements).
• Familiarity with digital assets, such as stablecoins, custody, or tokenization. This is useful context for where the platform is heading but not required.

Who thrives here

• Force multiplier. You measure your impact by how much faster everyone around you ships.
• Builder and teacher. You enjoy writing the tool and showing someone how to use it in equal measure.
• Pragmatic and curious. The tooling changes monthly. You try new things quickly, keep what works, and aren’t precious about the rest.
• Detail-oriented. AI makes "mostly right" cheap. You make sure we ship "right."
• Clear communicator. You write things down so they last: runbooks, specs, and decisions.

EvolutionsHQ is an equal opportunity employer. We welcome applicants of all backgrounds and make hiring decisions based on skills, experience, and fit for the role.