Home/Learn/Data & AI Engineer/30-day plan
30-day lab plan · 🤖 Data & AI Engineer

Become job-ready as adata & ai engineer in 30 days.

One focused task a day, four weeks, ending with a portfolio and a mock interview. Built for students and engineers in India who learn by doing — about 2 hours a day on weekdays, a little more on weekends.

What you needInterview questions →Role path →
Your kit

Set up before day 1

Python, DuckDB/PostgreSQL, dbt, Airflow (or Dagster), an LLM API key or local model (Ollama), and our LLM-choice and RAG guides. Everything here is free or has a free tier; a laptop with 16 GB RAM is enough for most labs (32 GB for nested clusters).

  • Daily ruleFinish the task, write 5 lines of notes, commit to Git. Notes become your interview stories.
  • Weekend ruleDay 7 of each week is a review: redo the hardest task from memory.
  • Stuck rule30 minutes max, then ask in a community or our engineer network.
Week 1

Data foundations

  1. Day 1SQL drills with window functions on a public dataset
  2. Day 2Pandas → DuckDB: profile and clean a messy CSV
  3. Day 3Dimensional model: build a star schema in PostgreSQL
  4. Day 4Parquet vs CSV benchmark; partitioning
  5. Day 5Airflow DAG: ingest → transform → load, with retries
  6. Day 6Data quality tests (dbt tests/Great Expectations)
  7. Day 7Review: data dictionary + lineage sketch
Week 2

Analytics & ML basics

  1. Day 8dbt project with staging/marts models and docs
  2. Day 9Metabase/Superset dashboard on your marts
  3. Day 10Train a classification model (scikit-learn); evaluate properly
  4. Day 11Feature engineering and cross-validation
  5. Day 12Model serving as an API (FastAPI) with logging
  6. Day 13Drift monitoring concept with a simple check
  7. Day 14Review: experiment log and results
Week 3

LLM applications

  1. Day 15Run a local model (Ollama) and a hosted API; compare cost/latency
  2. Day 16Build RAG over your own documents (pgvector); measure retrieval quality
  3. Day 17Prompt engineering + structured outputs (JSON schema)
  4. Day 18Evaluation set and automated judge; regression tests
  5. Day 19Guardrails: injection tests, output validation, PII masking
  6. Day 20Simple agent with 2 tools (search + calculator) and approval gate
  7. Day 21Review: write the responsible-AI checklist
Week 4

Platform & portfolio

  1. Day 22Streaming mini-lab: Kafka → consumer → ClickHouse/DuckDB
  2. Day 23Orchestrate the whole pipeline end to end; alerts on failure
  3. Day 24Cost model: GPU on-prem vs cloud for your inference load
  4. Day 25Data governance: catalog, access, retention (DPDP)
  5. Day 26Write a case study: problem → pipeline → model → impact
  6. Day 27Mock interview with our question bank
  7. Day 28Publish repo + dashboard screenshots; update resume
After day 30

What you will have

  • A portfolio repoDiagrams, configs, scripts and runbooks a hiring manager can read in five minutes.
  • Interview storiesReal faults you fixed — the answers to “tell me about a time…”.
  • A certification targetSee the role path for which exam to book next.
  • A place in our networkFinish the plan, share the repo, and apply — we mentor and place engineers who complete it.

Finished the plan? Show us.

Share your repo and notes; completed plans get a technical conversation with our architects and first paid tasks through the engineer network.

Apply to the networkPractice interview questions