AI Automations Product Engineer

Building, shipping, and supporting AI-powered tools and automations for 923 AI Automation clients. The role sits at the intersection of product thinking and AI engineering—closer to “product manager who builds confidently with modern AI development tools” than “AI research engineer.” You'll work with tools like Claude Code, the Anthropic API, and AI-augmented low‑code platforms to deliver real systems to real clients, end‑to‑end.

Enjoy this role if you:

• Like framing complex, ambiguous problems and working through them to clear solutions
• Enjoy ideating, building, and seeing things ship — then explaining clearly how and why they work
• Prefer direct interaction with a mentor over distant management
• Want autonomy and trust, not oversight for its own sake
• Want to understand why you're building something, not just what to build
• Are energized by seeing the whole picture—how a project connects to a client's business, a team's operations, and an emerging practice

Frustrating for you if:

• Prefer to work deep in code with minimal client or stakeholder contact—the role sits in the middle of client delivery, not behind it
• Are looking for a role with highly structured processes, detailed specifications, and predictable work—this department is actively shaping its own way of working, and the person in this seat will be part of that shaping

Requirements

Solutions Engineering & Client Delivery:

• Strategy & Growth Pattern Contribution—surfacing new approaches, tools, and workflows discovered during client delivery, and contributing them to the department's knowledge base and operating models. Ideas are validated on real client work before they become standard.
• Continuous R&D—spending dedicated, reported hours testing new AI tools, models, and frameworks relevant to active or upcoming client work. Findings are shared with the Practice Lead and the wider department.
• Practice Building—helping shape the way the department works as it scales—contributing to project templates, runbooks, evaluation approaches, and reusable components seeded from real delivered work.

Discovery & Shaping Support:

• Joining discovery calls and shaping sessions with the Practice Lead to translate vague business problems into viable AI‑driven solutions. Initially, you'll join these calls to learn the pattern; over time, you'll take more of them directly as direction and confidence grow.
• Technical Feasibility—evaluating client data, systems, and APIs to determine what's genuinely achievable within a proposed time budget. Flagging risks, rabbit holes, and integration blockers early, before commitments are made to the client.
• Demo Build‑Out—building working demos for prospective clients that demonstrate the proposed solution against their real problem, ready to be presented on proposal calls.
• Proposal & Scope Input—contributing technical detail to proposals and statements of work—architecture approach, milestones, success criteria—which the Practice Lead translates into client‑facing documents.
• Client Onboarding & Education—helping client teams understand and confidently use the systems we deliver. This includes walkthroughs, training sessions where relevant, and clear handover documentation.

Technical Execution:

• Project Setup & Architecture—standing up new client projects—repository, environments, tooling, third‑party accounts—following the department's conventions. Designing the architecture of individual projects in line with 923's engineering and deployment standards.
• System & Process Design—designing the end‑to‑end system: how data flows in, how AI components fit into the pipeline, where humans sit in the loop, what the output looks like, and how the user actually engages with it. Thinking in terms of the job the system does for the client, not just the components that make it up.
• Agentic Design & Trade‑offs—making considered decisions about when to use agentic approaches versus deterministic pipelines versus human‑in‑the‑loop steps. Understanding the trade‑offs and being able to explain them clearly to the Practice Lead and to clients.
• Hands‑on Build—writing the code that makes the system work, using AI development tools fluently. Python‑first stack, typically lightweight web frameworks (FastAPI, Gradio) and serverless runtimes (AWS Lambda). AI‑augmented practice is expected and welcomed—the role is defined by what ships, and real progress for our clients.
• Prompt Engineering & Model Selection—designing and refining the prompts, system instructions, and model choices that drive AI behaviour. Balancing accuracy, cost, latency, and privacy against the client's real constraints.
• Data Engineering—cleaning, structuring, and processing client data into a form the AI layer can work with. Building the ingestion and processing layers that feed the AI components.
• Deployment—working directly with 923's DevOps team to deploy services to client cloud accounts following 923's deployment standards—source control conventions, commit and PR protocols, secrets management, environment setup, and handover protocols.
• Validation Before Production—testing AI outputs against real client data before go‑live. Building the dry‑run and validation patterns that let clients sign off on output quality before we enable production writes. Post‑deployment observability is set up to 923 standards.

Project & Quality Management:

• Project Tracking—working from the department's Monday.com board, following established task, branch, and PR conventions. Tickets are written by the Practice Lead during onboarding; over time, you'll draft your own and the Practice Lead will review.
• Client Communication—joining and running technical portions of weekly client meetings. Initially alongside the Practice Lead and Account Manager; over time, running technical calls solo where the direction is clear. The Account Manager owns the relationship; you own technical clarity.
• Code Review—submitting clean PRs that pass pre‑commit checks and link to tracked work. Engaging in constructive, question‑based review with the Practice Lead and 923 engineering peers.
• Quality Assurance—testing AI outputs for failure modes, hallucinations, and edge‑case behaviour before anything reaches a client. Implementing the validation patterns the department uses to build client trust.
• Handover & Documentation—completing the delivery checklist before any project closes—documented architecture, credentials handover, runbook, walkthrough video, and third‑party account transfer to client billing.

Working Inside the Department:

• Tooling—using the department's stack—Monday.com for tasks, Notion for capture, Harvest for time, Slack for communication, Fathom for meeting intelligence, the department's Claude skill library for recurring workflows. Friction with the tooling is surfaced to the Practice Lead rather than worked around.
• Idea Flow—ideas, improvements, and concerns flow through the Practice Lead, who carries them forward to the wider 923 leadership team as appropriate.
• Documentation—contributing to the department's operating model and shared framework as patterns stabilise from real delivery.

Benefits:

• 100% remote job
• Annual paid vacation: 20 working days per year during the first 3 years, further increasing to 25 days in later years
• Paid sick leave and holidays
• Reimbursement of expenses for professional development courses/certifications (up to 100% in agreement with the Manager)
• Well‑being budget
• Team and atmosphere: we hire people we trust, so you'll be part of a strong positive engineering culture, a tightly‑knit team of professionals with a good sense of humor