The senior engineer who ships an agent makes 40% more. The PM who specs the eval runs the AI initiative. The platform lead who built the gateway leads the AI team. Become that person — in 8 weeks. For your current role, or your next interview.
We didn't build this for absolute beginners or AI researchers. We built it for the senior tech professional who already knows their craft and wants AI fluency layered on top.
You have an AI system design interview coming up — and no framework to prep with.
You read 12 AI tutorials a week and still feel behind.
Your team is shipping AI features — and you're the one who isn't.
Your job description has an AI section that didn't exist 18 months ago.
You want to lead AI work at your company, not ask questions in AI meetings.
You've shipped real software for 5+ years and want depth, not hype.
The engineer who ships a working agent in two days makes more than the one who can't — sometimes 40% more. The platform engineer who built the company's AI gateway is now running the AI platform team. The PM who can credibly spec an eval is the one leading the AI initiative.
The gap isn't closing. It compounds with every quarter.
And the path to that fluency is broken. Generic AI courses teach a little of everything, badly. Tutorials and blog posts are everywhere — but the half-life of AI best practice has dropped to weeks. What was sharp last quarter is mid this quarter.
Refactor4AI is built around the one thing that actually matters: becoming the most AI-fluent person at your company in the seat you already hold.
Not a researcher. Not a generalist. The person whose role is exactly what yours is — who can build, ship, and review AI work in your stack, in your meetings, in your code reviews.
We pick your role. We meet your skills where they actually are. We update the curriculum monthly so what you learn this week is what shipped this week.
You don't refactor your skills once. You refactor them continuously.
We built the curriculum backwards from two questions: what is a senior engineer, platform lead, or PM actually being asked in 2026 AI interviews — and what do they ship in their first 90 days on the job? Every module exists because the answer pointed to it.
Three role-specific tracks · 33 modules · 250+ lessons. Pick your role, scan the journey, click any module to see what's inside.
From AI-assisted development to shipping production agents with MCP, evals, and OWASP-aligned safety. Closes with case-driven AI system design. Calibrated for senior engineers who already ship code and need to layer AI fluency on top — including concrete patterns from the AWS Bedrock, Microsoft Foundry, and Google Vertex ecosystems.
How modern LLMs work — transformers, attention, MoE. Tokens and context windows. Reasoning vs base models (when to use o-series / Claude reasoning vs Haiku-class). 1M+ context tradeoffs. Multimodal capabilities. Cost & latency models. The capability map you need before writing any prompt.
All lessons free during early access — sign up to start.
Cursor, Claude Code, Windsurf, Devin, Copilot Workspace — when each wins. Effective prompting for code. Spec-driven development (the rising 2026 pattern). AI-generated PR review (blocking vs advisory). Refactoring at scale. When NOT to trust AI output.
All lessons free during early access — sign up to start.
Anthropic, OpenAI, Gemini APIs. Cloud SDKs (Bedrock, Foundry, Vertex). Structured outputs / JSON mode. Tool calling as the integration pattern. Streaming. Prompt caching (~90% cost cut on repeated prefixes). Batch processing. Multimodal inputs.
All lessons free during early access — sign up to start.
Embeddings (model selection in 2026). Chunking strategies — semantic, hierarchical, parent-document. Vector DBs (pgvector, Turbopuffer, Pinecone, Qdrant). Hybrid search (BM25 + vector + re-ranker). Query rewriting & decomposition. Long-context vs traditional RAG. Retrieval evaluation with RAGAS.
All lessons free during early access — sign up to start.
Model Context Protocol deep dive. Building MCP servers (TypeScript + Python). Consuming them across providers. Tool design ergonomics — descriptions matter more than names. Authentication (bearer tokens; OAuth for remote MCP). Governance and rate limits. The 5,800+ public server ecosystem.
All lessons free during early access — sign up to start.
Single-agent, multi-agent, supervisor patterns. Framework choice — LangGraph (production reliability), CrewAI (fast workflow), PydanticAI (type-safe). State and memory. Human-in-the-loop checkpoints. A2A protocol for cross-agent communication. Computer use, browser agents, voice agents.
All lessons free during early access — sign up to start.
Eval-driven development as a discipline. Building golden datasets (100+ cases). RAGAS metrics. LLM-as-judge with human calibration (sample 30+ pass cases manually). Binary PASS/FAIL > 1-5 ratings. Regression testing prompts in CI. Production observability (Langfuse / Langsmith / Braintrust). Hallucination detection. Drift monitoring.
All lessons free during early access — sign up to start.
Real-time speech-to-speech (Foundry Voice Live; Anthropic + ElevenLabs; OpenAI Realtime). Browser agents. Computer use (Anthropic Computer Use, OpenAI Operator). UX patterns for non-text AI. Latency budgets and how they shape architecture. Failure modes specific to voice — interruptions, partial transcripts.
All lessons free during early access — sign up to start.
Cost engineering and token economics. Prompt caching strategies (~90% savings). Semantic caching at the gateway. Model routing (cheap → expensive). Security: OWASP LLM Top 10 — prompt injection defense (constrain behaviour, segregate content, adversarial testing). Output validation. Distillation. Model gateways (LiteLLM, Portkey).
All lessons free during early access — sign up to start.
Designing systems AI-first vs bolt-on. Persistent memory and personalization. Conversational UX vs traditional UX. Failure modes and graceful degradation. Human-AI collaboration patterns (Klarna's reintroduction-of-humans lesson — hybrid > full automation). Reading the next 3 years.
All lessons free during early access — sign up to start.
Case-driven, interview-grade AI system design. Design a customer-support agent at fintech scale (RAG + tools + escalation + audit). Multi-tenant AI features for SaaS — per-tenant RAG isolation, cost attribution, noisy-neighbour problems. Agent orchestration patterns (queues, retries, idempotency, dead-letter). Real-time vs batch agentic systems. Multi-step agent capacity planning under cost ceilings. Failure modes you only see at scale — cascade hallucinations, eval drift, prompt regression. The exercise senior interviews actually test.
All lessons free during early access — sign up to start.
Multi-week project, full evals + observability, employer-style rubric, public artifact for your portfolio.
The 10-minute skill assessment routes you to the right starting point.
Every track ends with a portfolio-grade capstone. Multi-week, employer-style rubric, public artifact. The work you do in Refactor4AI is the work you walk into your next interview with.
Multi-week project where you pick an actual feature, write the spec, build the agent, wire MCP tools, ship evals, and deploy with observability. Employer-style rubric.
End-to-end build of the thing platform teams ship at scale: auth, rate limits, cost controls, prompt-injection defense, observability, self-service onboarding for product teams.
PRD + eval plan + GTM brief + EU AI Act risk classification for an actual AI feature. Built to be the thing you walk into a senior PM interview with.
Senior interviews at Anthropic, OpenAI, Google DeepMind, FAANG and AI-first startups now include AI system design as its own dedicated round. Refactor4AI is built backwards from those rubrics — so you walk in with the framework, the vocabulary, and a portfolio piece to talk about.
"Design a customer-support agent at fintech scale (RAG + tools + escalation + audit)."
"Design an AI code-review system for GitHub-scale. Walk me through cost, latency, evals, failure modes."
"Design the AI gateway for a 5,000-engineer company. Multi-tenant, cost-attributed, prompt-injection-defended."
"Your AI bill is $2M/year. Get it to $200K without losing quality. Talk me through it."
"Pick an AI feature and walk me through the PRD, eval plan, success metric, and EU AI Act risk classification."
"Pricing — your AI feature costs $0.40 per session. Per-token, per-action, or value-based? Defend."
Clarify, architect, AI-specific layers, non-deterministic concerns, scale & failure modes. The structure that lets you handle any AI design prompt without freezing.
Prompt caching. Model routing. OWASP LLM Top 10. RAGAS. LLM-as-judge. EU AI Act risk classification. Cost-attribution at multi-tenant scale. You'll know them all cold.
The capstone you ship through Refactor4AI becomes your talking point. Real working artifact, employer-style rubric, deep-dive doc — you don't show up empty-handed.
Module 11 in every track is built backwards from the FAANG / AI-lab hiring rubric. Or read the standalone interview-prep guide.
We're not the only way to learn AI. But for the senior tech professional with limited evenings, here's why we built Refactor4AI.
| YouTube tutorials | Generic AI course | Refactor4AI | |
|---|---|---|---|
| Role-specific (you don't see other roles' content) | |||
| Updated for what's shipping this quarter | |||
| Production patterns from real companies | |||
| Portfolio-grade capstone you can interview with | |||
| MCP, evals, agents, EU AI Act covered in depth | |||
| Sequenced learning path (not a random playlist) | |||
| Free | |||
| Time to ship something real | Years | 30+ hours scattered | 8 weeks structured |
We've tried each. We still recommend a great YouTube channel for any specific concept — but for "I want to be the AI person at my company" there isn't a substitute for a sequenced curriculum.
If your day-to-day still feels familiar, it's because the change is happening around you faster than your habits are catching up. These are the shifts our curriculum is calibrated to.
A small selection of what's happened in the last six months.
Six weeks of Refactor4AI and I shipped my first agent into production. I'm now the AI lead on the payments team. The curriculum caught gaps I didn't know I had — eval coverage was the thing that pushed me from prototype to production.
"I went from being the PM that asks questions in AI meetings to the one running them. Refactor4AI teaches AI as something you ship through a real product process."
"The platform-engineering track is the only curriculum I've found that takes the infra side of AI seriously. Cost engineering, prompt injection, EU AI Act — all in one place."
The whole platform — every track, every lesson, every deep dive, your public portfolio. No credit card. We'll add paid tiers when we know what people actually use.
Mid- to senior-level tech professionals — engineers, platform/DevOps, PMs — who already know their craft and need to layer AI fluency on top. If you've shipped real software and feel like AI is slipping past you, you're exactly who we built it for.
Three things. It's role-specific — engineers, platform people, and PMs all see different curricula. The curriculum is updated monthly because in AI, anything older than six months is outdated. And every module is anchored on a real shippable build, not abstract theory.
Updated monthly. The current version covers MCP, reasoning models, agent orchestration, the EU AI Act, eval-driven development, prompt caching, and AI cost engineering. We publish a public changelog so you can see what changed and why.
Module 1 of every track assumes zero AI background — it covers how modern LLMs actually work, capabilities, limits, and where the field is. You skip what you've already mastered.
Yes. The Team plan (when it launches) includes admin dashboards and skills reporting. We have a one-pager you can forward to L&D — request it via the team plan link.
The capstone is a real, shippable project that becomes the centerpiece of your portfolio. The structure is built so the work you do here is the work you can show in interviews.
Find your gaps. Build something real. Ship it on your resume.
Start free