$200/month · Billed quarterly · 30-day money-back

Learn data science with a group.

Not another course you'll abandon. Not solo YouTube rabbit holes. A small group of real people learning together — structured plan, weekly accountability, and a curriculum built from books, papers, and lectures we trust (no paid course subscriptions to chase).

Apply for the next cohort → See how it works
months to hired
18
career tracks
5
money-back
30-day
hrs/week
10–15

You don't need to be a math genius to start.

The only requirement is that you keep showing up. The group makes that part easier than you'd think.

Career changer

Switching from another field entirely

Marketing, finance, healthcare, teaching — you've built transferable analytical instincts and now you want to formalise them. We start from zero and move at a pace that respects your full-time life.

CS / STEM grad

Degree in hand, direction unclear

You know how to code but you've been staring at the ML landscape wondering where to focus. The curriculum and the group give you a path, structured resources, and people to learn alongside.

Working professional

Upskilling while staying employed

You have a job and a life. 10–15 hours a week is real but not punishing. The day-by-day pacing guide means you never spend two hours deciding what to do — you just open the guide and do the next thing.

Curious beginner

Never touched Python. That's fine.

The first three months are pure foundations — Python from scratch, math intuition with BetterExplained and 3Blue1Brown, and classical statistics. If you can follow a recipe, you can follow this.

A week in the group.

Four touchpoints a week. Each one is small enough to stick.

  • Mon

    Study session — book chapter or video

    Every day has a specific reading or video with an exact chapter reference. No ambiguity. No decision fatigue. Open the guide, do the thing.

  • Wed

    Hands-on coding or LeetCode

    Apply what you read. Build something small or solve a specific coding problem. Push it to GitHub. Employers hire evidence — this is how you build it.

  • Fri

    Essay or video essay — teach it

    Write 500 words or record 5 minutes explaining what you learned. Teaching is the fastest way to find your gaps. If you can explain it, you understand it.

  • Sun

    60-min group sync — together

    One member presents a paper or demos a project. Everyone shares wins and blockers. Next week's goals go on the board. Same time every week.

The 18-month arc

  1. Months 1–3

    Python Coding Track

    Variables, data structures, algorithms, OOP, NumPy, pandas, matplotlib. Exercism exercises and LeetCode patterns with day-by-day specifics.

  2. Months 1–4

    Foundations (all members)

    Math intuition, statistics, classical ML (regression, boosting, random forests), and deep learning fundamentals. Portfolio Project #1: end-to-end classification with XGBoost + SHAP.

  3. Months 5–12

    Specialise in your track

    ML Engineer, Data Scientist, Computer Vision, or GeoAI. Week-by-week pacing with exact book chapters, arXiv IDs, lecture numbers, and project specs.

  4. Months 13–18

    Job acquisition phase

    Portfolio polish, resume and LinkedIn, NeetCode 150, ML systems design practice, and mock interviews. Goal: 2–3 parallel final-round loops by Month 18.

One group. Five paths.

All five tracks share the same foundations. You start together, then specialise from Month 5.

ML Engineer

Build and deploy the systems that run models in production.

PyTorch CUDA C++ JAX MLOps DDIA Deployment

Data Scientist

Answer hard business questions with data. Design experiments that trust themselves.

Causal inference A/B testing RecSys Forecasting LLMs for DS

Computer Vision

Teach computers to see. Cameras, detection, segmentation, 3D, and beyond.

YOLO SAM ViT NeRF Diffusion OCR Tracking

GeoAI

Apply ML to maps, satellite imagery, and the physical world.

GIS Earth Engine Satellite ML torchgeo R spatial

NLP / LLMs

Build with language models — fine-tuning, RAG, agents, evaluation, and the systems behind them.

Transformers RAG Fine-tuning Agents Evals Tokenization Prompting

Sessions from the group.

A look at how we actually meet, present, and learn together.

Study Group — Session 1

Study Group — Session 2

Study Group — Live Session

The curriculum is built from free, open-access material.

No paid course subscriptions. No expensive bootcamps. The best free resources on the internet — books, papers, YouTube, and hands-on practice — curated and organised into a day-by-day pacing guide with exact chapter and section references. You pay for the cohort and the facilitation; the materials themselves are free for you to keep, forever.

Download the pacing guide

The full 18-month plan as an editable spreadsheet — five days a week, every track, every resource.

↓ Download .xlsx

Free textbooks

  • Deep Learning — Goodfellow et al.
  • ISLP — James et al.
  • Dive into Deep Learning
  • Think Python 2e
  • Causal Inference: The Mixtape
  • What If? — Hernán & Robins

YouTube courses

  • Karpathy — Zero to Hero
  • fast.ai — Practical DL
  • Stanford CS231n & CS224N
  • 3Blue1Brown — Calculus & LA
  • GPU MODE — CUDA lectures
  • StatQuest — full playlists

Must-read papers

  • Attention Is All You Need
  • ResNet, YOLO, SAM, BERT
  • Flash Attention
  • NeRF, 3D Gaussian Splatting
  • XGBoost, LightGBM, CatBoost
  • 60+ papers, graded

Coding practice

  • realpython.com
  • exercism.org — Python track
  • LeetCode — NeetCode 150
  • DataLemur — SQL practice
  • GitHub — weekly commits

Engineering blogs

  • Distill.pub
  • Jay Alammar's Blog
  • Netflix Tech Blog
  • Spotify Engineering
  • Colah's Blog
  • Chip Huyen's Blog

Pacing guide

  • 5 days × 5 tracks × 48 weeks
  • Exact chapter + section refs
  • arXiv IDs for every paper
  • Exercism exercise slugs
  • LeetCode problem numbers
  • Essay prompt each Friday

Your facilitator

Christina Bernard

Founder, Geeky Insights · Dallas–Fort Worth · LinkedIn ↗

I don't just teach data science — I've lived the journey you're about to take.

A former Deloitte Lead Data Scientist turned independent AI strategist, I've spent a decade building, deploying, and explaining ML systems across industries. More importantly, I know how to make technical concepts land — for students, for clients, and for executives who've never touched a line of code.

Why I'm the right facilitator for you:

  • Deep technical range — NLP, causal/uplift modeling, AWS MLOps pipelines (SageMaker, Docker, Lambda), A/B testing, and end-to-end ML pipeline redesign
  • Proven teacher — I've trained classrooms of 30 students and created technical onboarding docs that cut support tickets and accelerated new team ramp-up time
  • Built in public — I write a weekly AI newsletter, have published 10+ technical blog posts per month, and run a community for early-stage founders
  • I started where you are — no traditional DS degree; I built my career through self-directed learning including fast.ai, which means I know exactly where people get stuck
Connect on LinkedIn Read the newsletter

Honest answers to honest questions.

How much does it cost?

$200/month, billed quarterly ($600 every 3 months). You pay nothing to apply — billing only starts after you're accepted into a cohort. Every cohort comes with a 30-day money-back guarantee: if it isn't right for you in the first 30 days, we refund the full quarter, no questions asked. The curriculum itself is built from free, open-access material — no paid course subscriptions or extra fees.

I've never written a line of Python. Can I really do this?

Yes. The first three months are the Python Coding Track — day-by-day from "what is a variable" to NumPy, pandas, and basic algorithms. The math is introduced visually before any formal notation appears.

What is the weekly time commitment?

10–15 hours a week on your own schedule, plus one 60-minute group sync. Typical: 2–3 hrs Monday (reading), 2–3 hrs Wednesday (coding), 1.5 hrs Friday (essay), 1 hr Sunday sync.

What if I fall behind?

The pacing guide is a guide, not a law. If you fall behind by a week or two, you catch up. If something major comes up, you can rejoin the next cohort where you left off.

How big is the group?

Cohorts are 20–25 people. Big enough to keep momentum when life gets in the way for a few of you, small enough that you'll actually know everyone by Month 2. The weekly sync uses structured turn-taking so everyone gets a chance to speak.

Which track should I pick?

All five tracks share the same Foundations for the first four months. You don't have to decide until Month 5. Pick MLE for systems, DS for product/business, CV for cameras and images, GeoAI for maps and satellite data, and NLP/LLMs for language models, RAG, and agents.

Investment

One simple price. No upsells, no upgrade tiers, no "premium" version of the cohort.

Cohort tuition

$200 /month

Billed quarterly — $600 every 3 months

  • Day-by-day pacing guide for all 18 months
  • Weekly 60-minute live group sync, every week
  • All five tracks (ML Engineer, DS, CV, GeoAI, NLP/LLMs)
  • Curriculum built from free, open-access materials — nothing extra to buy
  • Direct facilitation from Christina — not a TA, not a community manager

30-day money-back guarantee

Show up for the first 30 days. If it isn't right for you, we refund the full quarter — no questions, no forms, no negotiation.

You pay nothing to apply. Billing only starts after you're accepted into a cohort.

Apply for the next cohort.

Small group. $200/month, billed quarterly. 30-day money-back. Tell us a bit about yourself and we'll be in touch about the next start date.

The "why" question is the most important field on this form. Real answers get real responses.

Contact & Basics

We currently operate in the United States only. Useful for scheduling the weekly sync across time zones.

Where You're Starting From

No "right" answer. Be honest — it helps us pace the group.

One line, in your own words.

How would you describe your current Python level? *

Do you have any coding experience outside of Python? *

JavaScript, R, SQL, C++, shell scripts — anything counts.

How would you describe your math comfort? *

Why You Want To Do This

This is the most important part of the form. People who write a real answer here are the ones who show up.

3–5 sentences. What's actually driving this for you right now?

One note: "I want a six-figure job" is not a reason — AI can already do the basics, and the entry rungs of that ladder are getting kicked out from under it. Real job security comes from depth of knowledge in a domain you actually understand. We want to hear what you're genuinely curious about.

What track are you most drawn to? *

No judgment. Knowing the pattern helps us help you not repeat it.

Commitment & Logistics

Can you genuinely commit 10–15 hours per week for the next 12 months? *

Moving, newborn, big project at work — better to know now than three weeks in.

Anything Else

Optional. Often the best answers come from here.

No payment to apply — billing only starts after you're accepted into a cohort. We'll be in touch about the next start date.

Having trouble submitting?
  • ERR_BLOCKED_BY_CLIENT — A browser extension or privacy tool blocked the form. Try a private/incognito window or another browser.
  • 502 error — Our server couldn't complete the submission. Please try again in a few minutes.

Geeky Insights. An 18-month study group for people who want to actually learn data science.
Questions? coaching@geekyinsights.com