THE ROLEAI training that sticks are run by someone who can do the work, not just talk about it. That is the bar for this role.
As an AI Training Engineer at Blank Metal, you split your time evenly between two jobs that make each other better. As a member of the training team, you are in front of clients, leading Claude Code and Cowork training. You own the outcome and the agenda. You are heads-down on Shippy, our internal platform, shipping features and fixes on a two-week cadence as part of the engineering team. The friction you watch clients hit in training becomes a backlog you fix in the product. The depth you build in the codebase is what makes you credible when a client CTO starts asking hard questions about MCP, RBAC, and platform readiness. You will coordinate with both the Head of Engineering and the Head of Training and Enablement to keep the two halves in balance.
Because you deliver real client work half the time, you are held to the same engineering bar as every other engineer here. The 50/50 is the steady state, not a discount on either half.
We are looking for candidates who are excited by the challenge of balancing training and engineering. This role is a great fit if you enjoy staying sharp with technical tools while helping others learn to use them effectively.
WHAT YOU GET TO DOLead the Room: Training & Enablement (50%)- Serve as lead trainer for Claude Code and Cowork sessions across in-person, virtual, and hybrid formats, owning session flow and outcomes for groups of 30 to 60
- Command a mixed room: field live questions, manage pacing, and hold authority whether the audience is individual contributors or the C-suite
- Troubleshoot participant setup live, including WSL, API authentication, SSL certificates, path variables, and enterprise IT restrictions, without losing the room
- Run technical advisory calls with client IT teams on security posture, RBAC, connector tiering, context architecture, and environment readiness
- Translate agentic AI for non-technical audiences: the difference between chat and agents, security tradeoffs, and governance, in plain language
Ship the Platform: Shippy Engineering (50%)- Ship features and bug fixes across the Shippy ecosystem (Blank Metal's internal AI-enabled operating system) on a two-week iteration cadence
- Pick up backlog assigned by the day-to-day engineering lead, keeping the queue moving whether the next task is an integration fix, a API bug, or net-new work
- Demonstrate working fluency across the stack: Next.js/React on Vercel, Neon/Postgres and the Databricks lakehouse, HubSpot, GCal, and Insperity integrations, Google SSO, etc.
- Hold opinions and make architecture recommendations with the technical architect, especially on feasibility tradeoffs, API contracts between repos, and how new products plug into the underlying code and infrastructure
- Operate as a contributor, not a gatekeeper. The team runs a "whole company can contribute PRs" posture, and your job is to keep things moving and high quality
Build the Bench: Curriculum, Certification & Mentorship- Support tiered training tracks (101 vs 201) and build placement criteria to route participants to the right level
- Build and evolve train-the-trainer programs, certifying internal client evangelists and Blank Metal support trainers to scale delivery beyond the core team
- Progress through Blank Metal's certification path (shadow, second chair, first chair) on the training side while growing Shippy ownership on the engineering side, so the role compounds rather than fractures
- Mentor incoming trainers through the same pipeline and pair with associate engineers on PR reviews, walkthroughs, and bug bashes
Close the Loop: Insight to Product- Turn friction observed during client engagements and training delivery into Shippy backlog items, so the enablement half is not a one-way pull
- Spot expansion opportunities during training (use cases, workflow gaps, organizational dynamics) and surface them with enough context to scope
- Stay current on Anthropic releases, MCP, and agent orchestration so your teaching and your code both reflect the platform's present state, not last quarter's playbook
WHAT WILL HELP YOU SUCCEED- You have shipped software to production and can walk us through what you built, what changed after launch, and how it performed
- You meet the same engineering bar as our delivery engineers: strong across a modern stack (TypeScript/React, a backend language, databases, APIs, and cloud), and comfortable owning features end to end
- AI-native fluency with Claude Code and Cowork: connectors, skills and plugins, MCP integrations, token and context management, and agent orchestration. You have built skills, plugins, or custom workflows
- You can teach. You have led workshops or enablement sessions, and you know the difference between a session that felt good and one that changed behavior
- You can hold a room of 30 to 60 people, technical and non-technical, and stay credible when a skeptical engineer or a client CTO pushes back
- You can talk to an executive and an engineer in the same meeting without losing either
- Willing to travel as needed (currently around 30% of engagements) to deliver on-site training and build client relationships in person
Nice to have- Experience in private equity, financial services, or healthcare, where most of our training happens
- A track record building or scaling a training or certification program
- Familiarity with our stack: Next.js/React on Vercel, Postgres, and Databricks