Engineering Team AI Training

Hands-on training tailored to your codebase and your team. Live demos, paired sessions on real PRs, and code reviews that meet your engineers where they are. We do not do hypothetical workshops on toy problems.

Session Status

a7f29c

● recording live demos · real PRs

~/select-interactive · pairing-session ● LIVE

$si pair --repo your-codebase --pr live

✓ sharing .cursor rules · CLAUDE.md patterns

✓ recording session · notes will follow

Format: 1:1 / 1:N
Sessions: live
Review: real PRs
Delivery: remote
Engineers Trained: 20+ ↑ across 8 teams

How We Train

Real codebases, real problems, real outcomes

01

Your repo

Every demo runs here

02

Your PRs

Live review, written feedback

03

Your stack

Patterns adapted to your tools

04

Your pace

Workshop, multi-week, or ongoing

Capabilities

What we cover

  1. Live Demonstrations

    We show every workflow in real time on a real codebase. No slides about how AI works in theory. Just engineers demonstrating actual patterns and answering questions as they come up.

  2. Hands-On Workshops

    Your team drives the agents on their own work while we coach. Mistakes happen. We stop, debug them together, and the team learns the patterns by doing them, not watching.

  3. Code Reviews

    We review real pull requests from your active sprints. Written, detailed feedback on AI-assisted output, security, performance, and architecture, held to the same standards we apply to our own code.

  4. Pair Programming

    Live screen-share pairing on the gnarly tickets (big refactors, unfamiliar libraries, or tricky agent loops). One-on-one sessions or small group formats.

  5. Custom Curricula

    For teams new to TypeScript, React, or modern testing, we build a phased curriculum that brings the team current without dropping production work. The training and the work happen in the same codebase.

  6. Skeptic-Friendly Sessions

    Some senior engineers are skeptical of AI tools, and rightly so. We do not try to convert anyone. We run live demos and let results, not rhetoric, do the talking.

The Engagement Arc

Six steps from kickoff to operating independently

A typical training engagement runs four to twelve weeks, depending on team size and depth. We start with a kickoff and live demos to set context, then move into hands-on training, ongoing code reviews, and pairing as needed. Monthly check-ins keep the patterns sticky after we step away.

Training Formats

One-Day Workshop

A focused intro for teams that want to see agentic AI in action before committing to a longer program.

  • Format: 1 day · remote or onsite
  • Best for: Teams of 3–15 engineers exploring agentic AI for the first time.
  • Live demos on your codebase
  • Tool walkthroughs: Cursor, Claude Code, Linear
  • Q&A and follow-up notes
  • Recommended next steps tailored to your stack

Multi-Week Training

Most popular Structured training rolled out across multiple sessions so the team can absorb each pattern before the next one lands.

  • Format: 2-4 weeks · weekly sessions
  • Best for: Teams ready to commit to a coordinated rollout across multiple projects.
  • Weekly live demos and exercises
  • Code reviews on real PRs between sessions
  • Pairing sessions on demand
  • Final knowledge handoff and runbook

Ongoing Retainer

Continuous training partnership with monthly check-ins, code reviews, and new-tool evaluations as the AI ecosystem evolves.

  • Format: Monthly · remote
  • Best for: Teams that want a long-term partner to keep them current and accountable.
  • Monthly strategy session with engineering leads
  • Code reviews against your active sprints
  • New-tool evaluations and migration guidance
  • Direct Slack/Linear access for questions

Tools We Train On

  • Biome / Ultracite
  • Claude Code
  • Cursor
  • GitHub
  • Linear
  • Shadcn/ui
  • Tailwind
  • TanStack Start / Router / Query / Form / AI
  • TypeScript
  • Vite
  • Vitest

How We Work

01

Kickoff Call

We meet your team, learn the codebase, and identify what each engineer wants to get better at. We tailor the training to the gap, not a generic curriculum.

02

Live Demo Session

A single live session where we demonstrate the full agentic workflow on your repository. The team watches, asks questions, and sees the patterns end to end.

03

Hands-On Training

Your engineers drive the agents on their own work. We coach in real time, stop on real mistakes, and build the team's comfort with the patterns through practice.

04

Code Reviews

We review real pull requests from your sprints, written feedback delivered in your repository. Each review is a continuation of the training, on the team's actual work.

05

Pairing Sessions

On-demand live screen-share pairing for the harder tickets. Senior engineers from our team work directly with yours on real challenges.

06

Ongoing Check-Ins

Monthly sessions to address new tools, revisit progress, and keep the team current as the agentic AI ecosystem evolves. Optional after the initial engagement.

FAQ

Is training delivered in-person or remote?

Both. Most engagements are fully remote (we are based in Fort Worth and have run trainings for teams across the U.S.). Onsite is available for local clients and for major kickoffs.

How big a team can you train at once?

Workshops scale to about 15 engineers in a single session. Beyond that, we recommend splitting into smaller groups so every engineer gets hands-on time. We have run trainings for teams as small as 3 and as large as 50 (in cohorts).

What if some engineers are senior and some are junior?

Mixed-experience teams are the norm and they actually train better together. Senior engineers ask the architectural questions; juniors absorb the patterns faster. Both groups benefit from seeing how the other reasons through agentic workflows.

Do you provide recordings or just live sessions?

Live sessions are live (no replays of canned content). For onboarding new team members later, we provide custom-recorded walkthroughs of your specific setup as part of the final handoff.

What outcomes can we expect?

Most teams see measurable PR throughput gains within the first sprint after training. The compounding gains from running multiple agents in parallel, designing custom loops, and tightening review cycles typically take six to twelve weeks to settle in.