Joinloop Subjects: What You Can Learn Online in 2026

Joinloop runs live, cohort-based programs focused on one thing: building and shipping real AI systems. There are two tracks — one for engineers, one for product managers — plus a set of free webinars for people deciding where to start. This post lays out what's covered in each, who it fits, and how to pick.

The two tracks at a glance

	AI Engineering Cohort	AI for PMs Cohort
Who it's for	Engineers who write code and want to build agentic systems	PMs and aspiring AI PMs who own AI features
Length	7 weeks, 14 live sessions	8 weeks, live sessions + panel defense
You leave with	Deployed capstone + portfolio projects	End-to-end AI feature: PRD, prototype, eval framework, ROI model
Prerequisite	Comfortable in Python; some backend experience	Product experience; no coding required
Format	Live, weekend sessions, recorded	Live, weekend sessions, recorded

Both are taught by practitioners currently working in the field, both are project-first, and both end with you presenting something you built.

AI Engineering Cohort: the syllabus

Seven weeks, each building on the last like assembling a production system.

Week 1 — Foundations: GenAI foundations and prompt engineering; tool calling, MCP, and AI workflows.
Week 2 — Agents: Introduction to agentic AI; multi-agent systems and agent-to-agent (A2A) communication.
Week 3 — Retrieval: RAG fundamentals; memory and agentic RAG.
Week 4 — Applied: Building data agents (natural-language-to-database); RAG evaluation and AI quality systems.
Week 5 — Production: AI guardrails and security; production deployment of AI systems.
Week 6 — Design + build: AI system design masterclass; capstone build workshop.
Week 7 — Ship: Final capstone presentations; resume, LinkedIn, and AI career positioning.

Skills you walk away with: designing agent loops, wiring tools and MCP, building and evaluating RAG pipelines, adding guardrails, deploying to production, and reasoning about AI system design in a review.

AI for PMs Cohort: the syllabus

Eight weeks, structured around the decisions an AI PM actually makes.

Week 1 — Foundations: Predictive vs generative, specialist vs general; the ML literacy a PM needs; how the AI PM mindset differs from classical software.
Week 2 — The engine room: How LLMs work (tokens, context windows, temperature, the autoregressive loop); why models hallucinate; prompt engineering techniques.
Week 3 — LLM deep dive: The full model lifecycle (pre-training → fine-tuning → RLHF → eval → deployment); when to fine-tune vs prompt + RAG; the cost/latency/quality triangle.
Week 4 — Deciding and designing: The four-question AI fitment test; RAG deconstructed; defending an architecture choice in a design review.
Week 5 — Building: AI-native UX patterns and the trust gradient; hallucination handling; agents demystified (planner-dispatcher, tool use, memory, human-in-the-loop).
Week 6 — Measuring: Offline vs online vs shadow evals; LLM-as-judge and its failure modes; guardrails and the eval-driven improvement loop.
Week 7 — Shipping: AI feature cost models; launch playbooks (staged rollout, fallback, monitoring); the six AI PM interview question types.
Week 8 — Capstone: Team presentation (problem, fitment, architecture, prototype demo, eval results, ROI) and a panel defense where instructors challenge every choice.

Skills you walk away with: judging where AI fits, writing PRDs for probabilistic features, designing evals, modelling cost and ROI, and defending an AI product decision under scrutiny.

Free webinars: where to start if you're unsure

Joinloop runs standalone sessions you can attend without enrolling in a cohort. Recent topics:

Building AI agents that act, not just answer
Building a natural-language database agent in ~60 minutes
What the AI PM role actually demands now
Giving an AI memory that works
Claude Code from zero to power user
The four questions every AI agent must answer

These are a low-cost way to sample the teaching style and figure out which track fits.

Which track fits you?

Your situation	Track
You write code daily and want to build agents/RAG/multi-agent systems	AI Engineering
You're a backend/full-stack engineer moving into AI work	AI Engineering
You own or want to own AI features but don't code	AI for PMs
You're a PM who keeps getting blocked in AI design reviews	AI for PMs
You're an engineer who wants the product-decision lens, not the build	AI for PMs
You're not sure AI is your direction yet	Start with a free webinar

Beginner to pro: Both tracks assume you're not starting from zero in your own domain — an engineer should know Python, a PM should have shipped products. Within that, the cohorts run from fundamentals to advanced in the same seven or eight weeks, so a mid-career professional isn't sitting through an intro they don't need for long.

Common pitfalls when choosing

Picking the engineering track to "learn to code." It's for people who already code and want to build AI systems, not a first programming course.
PMs skipping the fitment and eval weeks. Those are the weeks that change how you work. Don't treat them as optional.
Enrolling without checking the schedule. Both tracks are live and weekend-based. Confirm the slots work for the full duration.
Ignoring the free webinars. They exist precisely so you can de-risk the decision. Use them.

Next steps

Read the full syllabus and schedule for your likely track: AI Engineering or AI for PMs.
Attend one free webinar to sample the teaching.
Block the cohort schedule in your calendar before enrolling.
If you're still between tracks, pick based on whether you want to build the system or decide and own it.

Both Joinloop tracks are live, small-batch, and project-first — you finish with something you built and can defend. Compare the AI Engineering Cohort and the AI for PMs Cohort, or see how to choose a credible online course.