Products
Growth engineering
No more boring admin
Working with agents
In your life
Making your company AI‑native
I help companies change how they work with AI: engineering teams moving to parallel agent work, and the rest of the organisation finding where AI actually helps and where it doesn’t.
That means 1:1 advisory with the people leading the change, next to hands‑on work in the codebase: how to upskill engineers and non‑engineers, which bottlenecks matter once developer speed no longer does, and what works in practice rather than in a demo. I have coached nine teams so far, from AI‑native engineering teams at a European cybersecurity company to advisory with the leaders of others.
If you’d like to talk about any of it, write to me at [email protected]. The free Augmented Mind Course shows how I teach it.
Bio
Fredrik Wollsén coaches companies through becoming AI‑native. He has coached nine teams so far, engineers and non‑engineers, pairing 1:1 advisory for the people leading the change with hands‑on work beside the team. He has been a professional software engineer for 22 years, the last decade making AI genuinely useful, and works with 10–15 Claude Code sessions in parallel every day.
Today he’s AI Engineer Principal Tech Lead at a European cybersecurity company, leading a team that’s been working AI‑native since August 2025, and mentoring more of its teams. Outside it, he coaches two more teams: at a non‑profit and at an investment fund. Previously Claude Code Coach at DNV Cyber (2024–April 2026), and before that Mozilla Research (web measurement).
All of it ships under Augmented Mind AB — his one‑person software company in Helsinki, Finland (business ID 3092558‑1). One human, a lot of agents.
Now
Coaching & engineering leadership
-
AI‑native transformation coaching: nine teams so far. The newest, from October 2026, is shaped nothing like a product team, which makes it a new test of what AI‑native means.
-
AI Engineer Principal Tech Lead at a European cybersecurity company, leading a 4‑engineer team that’s been operating AI‑native since August 2025. In June 2026 the team merged 844 pull requests against shipping product, 490 of them mine, with each engineer running 5–10 parallel Claude Code work streams.
Shipping
-
Augmented Mind Course (Walk → Run for engineering teams, free and self‑paced) at augmentedmind.io, and its sister AgentHerder Course (Run → Fly for individual engineers scaling to 10+ parallel sessions) at agentherder.com.
Writing
I write about AI‑native engineering, parallel agent practice, and what happens to software when dev speed stops being the bottleneck — at Fred Writes Words. The ones about AI‑native work are also here.
-
My coding agents would cost $31,000 a month if paid per token — seven weeks of my own Claude Code usage, measured the way OpenAI measured its researchers, and what it means for how companies should pay for AI.
-
This is getting ridiculous: I shipped 490 pull requests in June — the Playbook follow‑up, four months later: throughput numbers for every stage (Crawl 5–10 → Fly 100–200+ merged PRs per engineer per month), the per‑PR audit behind the count, and why the value multiple is three to four rather than ten. Flying, but in a greenhouse.
-
Local LLM speed on a Mac: Apple Silicon vs Intel — tokens per second on an M1 Max and an Intel i9, ollama vs SwiftLM up to 64K context, Whisper and vision models, and which model fits each RAM tier. More in Research.
-
The AI‑Native Engineering Playbook — how I actually work with 10–15 Claude Code sessions in parallel.
-
Augmented Mind Course — the full six-week AI‑native curriculum for engineering teams, free and self-paced.
-
AgentHerder Course — the free Run → Fly course for individual engineers already running parallel Claude Code sessions and scaling to 10+ sustained. Sister course to the AM curriculum.
Past work
-
Panelist, GOSIM AI Vision Forum Paris 2026 — Panel 1, “Agentic AI Systems: Human–AI Symbiosis.”
-
Claude Code Coach — coached engineering teams on AI‑native software engineering practices. Engagement concluded April 2026.
-
Web measurement engineer. Co‑author of the WWW 2020 peer‑reviewed paper The Representativeness of Automated Web Crawls as a Surrogate for Human Browsing (WWW ’20, ACM).
-
Independent and startup engineering work across data‑for‑good (Gapminder), adtech (Adoveo / NeamTime), and indie fintech — the first incarnations of Clerk.AI, which began in a 2016 accelerator and first shipped commercially in 2017. Bridge between the pre‑LLM era and where I am now.