Your definitive roadmap from zero to career-ready - for Data Scientists, ML Engineers & AI Engineers
"The best time to start was yesterday. The second best time is now."
3 Career Tracks | 100+ Curated Resources | 50+ Hands-On Projects | 100 Hackathon Ideas | 10 Deep Dive Blueprints | 12 Paper Digests
| Addition | Description |
|---|---|
| 10_Hackathons/ | 100 research-grounded hackathon ideas across 3 themes (Prolonged Coordination, Multi-Agent Collaboration, Adaptive Retrieval). All anchored on MongoDB Atlas + AWS Bedrock. |
| Deep Dive Blueprints | 10 full implementation blueprints with architecture diagrams, MongoDB schemas, 90-second demo scripts, and runnable Python starter code. Includes Viral Autopsy, Portfall, Carbon Lie Detector, and 7 more. |
| Starter Code | 4 runnable Python skeletons: ReasoningBank (LangGraph + MongoDB), A2A handshake, ColPali indexing pipeline, GraphRAG + HippoRAG PPR. |
| 11_Recent_Topics/ | 2025–2026 Trend Tracker (25 hot topics with maturity ratings), Stack Watch (MongoDB, AWS Bedrock, 13 LLM models), and 12 paper digests. |
| Paper Digests | 5-minute digests of the 12 most hackathon-relevant papers: ReasoningBank, Zep, SagaLLM, A2A, MCP, ColPali, Search-R1, GraphRAG, HippoRAG, and more. |
| 09_Competitions/ | Full competitive ML guide: Kaggle deep-dive, platform comparison, winning strategies, 20 landmark competitions. |
The internet is full of scattered tutorials. This repo is different:
- Structured pathways - not random links, but a curated journey from beginner to expert
- Role-specific - tailored tracks for 3 distinct, high-demand careers
- Hands-on first - learn by building real projects, not just watching videos
- Industry-aligned - skills and stacks that companies actually hire for (2025-2026)
- Free - 95% of resources linked are free or open-source
- Autonomously Updated - Managed by AI to ensure the content never goes stale
| You Are | You Should Go To |
|---|---|
| Curious beginner, no CS background | Start Here → |
| CS student exploring career options | Role Comparison → |
| Want a visual step-by-step roadmap | Learning Roadmaps → |
| Aspiring Data Scientist | DS Track → |
| Aspiring ML Engineer | MLE Track → |
| Aspiring AI Engineer | AIE Track → |
| Preparing for interviews | Interview Prep → |
| Looking for projects to build | Hands-On Labs → |
| Preparing for Kaggle / competitions | Competitions → |
| Brainstorming for a Hackathon | Hackathons → |
| Want the bleeding-edge trends | Recent Topics → |
| Negotiating a salary | Salary Guide → |
┌─────────────────────────────────────────────────────────────────────┐
│ │
│ DATA SCIENTIST ML ENGINEER AI ENGINEER │
│ ───────────── ─────────── ─────────── │
│ │
│ "What does "How do we "How do we │
│ the data say?" build & ship build products │
│ ML systems?" with AI?" │
│ │
│ Statistics + Software Eng + LLMs + APIs + │
│ ML + Insights MLOps + Scale Agents + UX │
│ │
│ [Explore →] [Explore →] [Explore →] │
└─────────────────────────────────────────────────────────────────────┘
Not sure which to choose? Read the in-depth comparison →
"Turn data into decisions"
| Phase | Topics | Duration (self-paced) |
|---|---|---|
| Beginner | Python, SQL, Statistics, EDA | 2-3 months |
| Intermediate | ML Algorithms, Feature Engineering, Visualization | 3-4 months |
| Advanced | Deep Learning, NLP, Time Series, A/B Testing | 4-6 months |
| Projects | End-to-end capstone projects | Ongoing |
"Build systems that learn at scale"
| Phase | Topics | Duration (self-paced) |
|---|---|---|
| Beginner | Python, Git, Docker, ML Basics | 2-3 months |
| Intermediate | MLflow, Feature Stores, Model APIs | 3-4 months |
| Advanced | Kubernetes, Spark, LLM Fine-tuning, Ray | 4-6 months |
| Projects | Production ML pipelines | Ongoing |
"Build intelligent products with AI"
| Phase | Topics | Duration (self-paced) |
|---|---|---|
| Beginner | Python, APIs, Prompt Engineering | 1-2 months |
| Intermediate | RAG, LangChain, Vector DBs, Agents | 2-3 months |
| Advanced | Fine-tuning, Multi-agent systems, Evals | 3-5 months |
| Projects | AI-powered apps & agents | Ongoing |
Before diving into any track, build these foundations:
- Mathematics & Statistics - Linear algebra, calculus, probability
- Python Programming - From syntax to software engineering
- Statistics for DS/ML - Hypothesis testing, distributions, Bayesian thinking
Not sure where you fit? Check the Learning Roadmaps → for Mermaid flowcharts with time estimates for each track.
Theory needs practice. We provide comprehensive guides for both competitive ML and rapid-prototyping events:
Hackathon Ideas Hub — 100 ideas, 3 themes, 10 blueprints:
| Theme | Focus | Ideas | Key Stack |
|---|---|---|---|
| Theme 1: Prolonged Coordination | Durable agents surviving weeks/months | 33 ideas | LangGraph + MongoDB checkpointer + EventBridge |
| Theme 2: Multi-Agent Collaboration | A2A protocol, EVINCE debate, agent swarms | 44 ideas | A2A + MongoDB Change Streams + Bedrock |
| Theme 3: Adaptive Retrieval | Self-improving RAG with bandit routing | 23 ideas | ColPali + HippoRAG + Voyage rerank-2.5 |
- 10 Deep Dive Blueprints — Full architecture, MongoDB schemas, 90-second demo scripts, build order, judging rubric alignment
- Starter Code — 4 runnable Python skeletons to get to a working demo in < 2 hours
- Competitions Overview — Kaggle deep-dive, platform comparison, winning strategies with code
Learning by doing is 10x more effective than passive consumption.
- Interactive Notebooks - Jupyter notebooks with exercises and solutions
- Mini-Projects - 1-3 day projects to cement concepts
- Capstone Projects - Portfolio-worthy end-to-end projects
Keep up with the bleeding edge. AI/ML moves fast, and this section tracks what is actually viable for production vs. what is still in the lab.
- 2025–2026 Trend Tracker - Emerging paradigms like Reasoning Models (o1, DeepSeek-R1), SLMs, and LLMOps.
- Stack Watch - The infrastructure and tooling movements that matter.
- Paper Digests - 12 digests of landmark 2024–2025 papers: ReasoningBank, A2A, MCP, ColPali, Search-R1, GraphRAG, HippoRAG, and more.
This repository is kept fresh, well-organized, and continuously improving by RepoSentinel - an autonomous multi-agent system.
Instead of relying solely on manual updates, RepoSentinel runs a continuous pipeline to:
- Analyze the repo against a curated DS/ML/AI topic taxonomy to identify content gaps and stale material.
- Scout the web (arXiv, YouTube, GitHub, Tavily) for high-quality, up-to-date resources.
- Synthesize updates by merging new findings directly into our existing Markdown files.
- Validate via a strict 5-check quality gate (formatting, link validity, syntax, deduplication, and completeness).
- Publish approved changes autonomously via Pull Requests.
Thanks to RepoSentinel's cross-run memory and self-improvement loop, this Learning Hub evolves at the speed of the AI industry itself.
- Salary Guide 2025/2026 — Compensation by role, level, location
- Interview Prep — 200+ questions with answers by role
- Portfolio Guide — How to stand out to hiring managers
- Resume Templates — ATS-friendly templates per role
- Networking Guide — How to land the job without applying blindly
- Books — The definitive reading list (free PDFs where available)
- Courses — Best free and paid courses ranked
- Research Papers — Landmark papers every practitioner should read
- Datasets — 50+ datasets for practice projects
- Tools & Stack — The modern AI/ML/DS toolbox
- Communities — Discord, Slack, forums, and Meetups
Use the skill badge system to track your journey:
| Badge | Level | Criteria |
|---|---|---|
| Novice | Just starting | Completed beginner section + 1 project |
| Practitioner | Building skills | Completed intermediate + 3 projects |
| Professional | Job-ready | Completed advanced + portfolio + interview ready |
| Expert | Senior-level | Contributing to open source, mentoring others |
Tip: Fork this repo and check off completed items in each section's README!
- Read the Role Comparison to pick your track
- Follow the roadmap for your chosen track (beginner → intermediate → advanced)
- Build at least one project per phase — add them to your GitHub
- Practice with interview prep questions as you go
- Contribute your learnings back (see Contributing)
This is a living document. Contributions are warmly welcomed alongside RepoSentinel's automated updates!
- Found a broken link? Open an issue.
- Have a better resource? Submit a PR.
- Want to add a project? See contributing guide.
MIT License - free to use, share, and build upon with attribution.
If this repo helped you, please give it a ⭐ - it helps others find it!
A massive thank you to everyone who has contributed to this learning hub!
Made with passion for the learning community | Updated May 2026