Private consultation · iOS × AI agents

Make your iOS repo usable by agents.

If Claude Code, Cursor, or Xcode agents feel impressive in demos but clumsy inside your real Swift codebase, the problem is rarely just the model.

Agents need a repo they can understand, commands they can run, feedback they can see, and a review loop that catches the dumb stuff before it reaches main. Most iOS projects were not built with that loop in mind.

Agentic iOS consultation

A private working session for senior iOS developers, indie app builders, and mobile teams who want practical help making AI coding agents useful on a real Swift project.

Private session Repo teardown Swift / SwiftUI / TCA Claude Code / Cursor / Xcode agents
Send me the problem → Ask about a team audit →

Useful if you are trying to

  • Set up Claude Code, Cursor, or Xcode agents on a real iOS repo.
  • Write project instructions agents will actually follow.
  • Make build, test, simulator, and preview commands agent-runnable.
  • Find why agents keep making wrong changes in your Swift codebase.
  • Design a safer human review loop for AI-generated changes.
  • Figure out whether your repo is agent-ready or agent-hostile.

What I can look at

Repo structure

Module boundaries, TCA seams, dependency shape, naming, and how much context an agent has to infer.

Agent context

CLAUDE.md, AGENTS.md, examples, conventions, constraints, and failure rules.

Feedback loop

Build, test, simulator, SwiftUI previews, screenshots, and how quickly bad work becomes visible.

Verification

Tests, snapshots, deterministic checks, review prompts, and rollback paths.

Shipping workflow

How an AI-assisted change moves from idea to branch to review to App Store-safe release.

Possible session shapes

One-off consultation

  • You send the repo or describe the failure mode.
  • We identify the bottleneck.
  • You leave with a concrete next move, not a lecture.

Repo teardown

  • I review the agent-readiness of the codebase.
  • You get findings, risks, and suggested changes.
  • Good fit before a team rollout or deeper sprint.

Team audit

  • For mobile teams testing agentic development.
  • Focus on workflow, safety, code review, and adoption.
  • Useful before standardizing tools across a team.

Small lab, if there is demand

  • If several senior iOS devs want the same help.
  • Small group, hands-on, no course platform cosplay.
  • Built from real problems people bring in.

Why me

I’m Igor Tarasenko, ex-Uber, lead iOS engineer in Amsterdam. I built WhisperBoard, an on-device transcription app with 50k+ downloads, a 4.8 App Store rating, and 1,035 GitHub stars. I’ve been running agent-driven iOS workflows before Xcode shipped native agent support.

I built preview tooling because the feedback loop was the bottleneck. That is the point here: make agents useful by giving them a repo they can understand, commands they can run, and evidence they can check.

Send me the messy version

You do not need a polished brief. Send the repo shape, the tool you are trying, what keeps failing, and what you want the agent to handle. I’ll tell you whether I can help and what the smallest useful session should be.

Email Igor →