Founding Engineering Manager / Lead AI Engineer

Philippines. Location: RemoteType: Full-time - founding leadership hireTeam: Leads a small, senior AI engineering team (2-3 today, growing)Growth path: -> Head of Engineering as the product scales to hundreds of tenants About the Agile & CoAgile & Co builds marketing and sales tools for home services and blue-collar businesses. We move fast, test ideas quickly, and care about software that creates real business results.

This is a founding-level leadership role for someone who can both run a team and build the system. You set the technical standard, run the engineering, and make the calls that decide whether our product scales cleanly to hundreds of business customers. We value skill, malleability, speed, and ownership - in a leader who still ships.

The RoleYou will lead the engineering of our flagship product: an autonomous AI marketing operative (internally, Emjae), sold as a multi-tenant SaaS. It doesn't just answer questions - it pulls data, builds reports, manages campaigns, and takes action across eight marketing platforms, all behind human approval. You will manage the engineering team, own the roadmap and delivery, and architect and ship the core of the product.

This is a greenfield build with a detailed technical spec already in hand - genuine agent autonomy, not a chatbot wrapper, with paying customers ahead of it. This is a deeply technical, hands-on leadership role. People management is the job, not a side quest - and so is living in the architecture and the code, setting the bar by example rather than from the standup. You lead because you're one of the strongest engineers in the room, not instead of being one.

What You'll OwnLeading the team and deliveryManage the engineers: hiring, onboarding, 1:1s, performance, and growth. You own whether the team ships. Run the project: roadmap, planning, delivery cadence, and accountability for milestones - translating the spec into a sequenced, shippable plan. Set the technical standard: code review, architecture decisions, and the engineering practices the team builds on. You are the technical voice in product decisions.

Building The ProductThe agent runtime - a persistent, code-executing AI agent with a skill-based memory system that runs work autonomously on a schedule, behind a draft/approval safety layer. Eight platform integrations - Google Ads, Meta Ads, WhatConverts, WordPress, Google Analytics, Google Workspace, Screaming Frog, and Google Search Console - built as typed, self-healing wrappers with full CRUD and draft-gated writes.

The infrastructure substrate - multi-tenant isolation, a warm-container sandbox pool on persistent storage, a job queue and worker pool, horizontal scaling, and a streamed frontend with server-side job lifecycle. Cost as an engineering responsibility - your architecture decisions are the margin. You own the model-tiering strategy, prompt caching, and batch routing that keep per-tenant cost healthy as the product scales.

Correctness and safety - an evaluation harness, prompt-injection defenses, versioned and rollbackable agent memory, and the human-in-the-loop approval system that gates every write. Alongside the platform, the team also ships the quick wins Agile & Co runs on - lead magnets, sales tools, dashboards, and report generators for home services businesses. You decide what's worth building now and what can wait, and you make sure both the platform and the small tools actually ship.

jsReact, TypeScript, and VitePostgreSQL or SupabaseClaude, OpenAI, Gemini, and open-source models (Llama, Kimi, Qwen, and similar), via OpenRouter and Fireworks (routed across cheap to frontier tiers)Google APIs, SEO data, Google Ads data, Exa for web search, scraping/crawling tools, CRM APIsHeadless browser automation (Playwright / Chromium)Job queues and workers (Celery or BullMQ), RedisVercel, Railway, and CloudflareLinux and SSHLogging and observability: Loguru, Python logging, Grafana, Loki, SentryClaude Code, Codex, Cursor, GitMCP, RAG, Hermes, and OpenClaw where they fit the project What We're Looking ForMUST HAVE:Engineering management and project leadership: you've managed engineers and run delivery - planning, roadmap, cadence, and accountability - not only mentored on the side; this role genuinely leads people.

Built and operated production systems: you've owned something real that ran reliably under load - production experience, not only prototypes. LLM application engineering: agentic loops, tool/function calling, prompt and context engineering, and a working grasp of how to make model-driven systems correct and predictable. Strong Python and TypeScript, and comfort across a modern web stack (React + Vite on the frontend).

Backend and infrastructure: job queues, container orchestration, persistent storage, and API integration at scale with proper rate-limit and retry handling. Cost-awareness as instinct: in an LLM product, model and routing choices are P&L decisions, and you architect accordingly. STRONGLY PREFERRED - this is where the product lives or diesAI safety and reliability instincts: prompt-injection defense, evaluation / golden-test harnesses, and human-in-the-loop design.

Hands-on with LLM cost levers: model tiering, prompt caching, batch processing. Multi-tenant SaaS architecture and the isolation and security it demands. Marketing-platform APIs (Google Ads, Meta) are a plus, not a requirement - the domain is learnable; the engineering and leadership judgment is not. How You WorkSound judgment under ambiguity - much of the spec is "recommended default, but the team decides"; you'll live in those decisions and own the outcome.

You'd rather be right than look right - you change course quickly when the evidence says you're wrong. You finish things - tests, edge cases, error handling, and the unglamorous operational work that keeps a system alive. You move fast and ship, and you're self-directed - you set your own priorities. You raise the bar for the people around you What This Role Is NotA pure people-manager who never touches code - or a senior IC with no interest in leading a team.

A research-only role. A heavy-process role. A role for someone who needs perfect specs before starting - much of this is yours to decide. A thin AI-assistant role - this is genuine agent autonomy, with real platform access and real consequences. How To ApplyIMPORTANT: include at least one specific example of something you have built and shipped - and, if you've led, one example of a team or delivery you have owned.

Strong applications will include:The problem, and what you built or led (with the rough stack and the impact)If you led it: how you ran the team or made the call under ambiguityOne link or artifact if you have one - repo, demo, writeupWe care most about what you have shipped, the teams and deliveries you have owned, how fast you learn, and how you handle unclear problems.