What will I learn?
- Gemini API
- LangChain
- LlamaIndex
- pgvector / Pinecone
- RAG Pipelines
- Function Calling
- Evals & Observability
- Python
What do I need before starting?
- Python proficiency
- Basic understanding of machine learning concepts
What is the weekly roadmap?
Week 1: LLM Foundations & Prompt Engineering
Model architectures · Tokenization & Context Windows · Few-shot & Chain-of-Thought prompting · Structured JSON Outputs
Week 2: Gemini API & Multimodal Inference
Gemini 1.5 Pro & Flash · Audio, Video & Image parsing · Streaming responses · System instructions
Week 3: Vector Databases & Embeddings
Text embeddings theory · pgvector in PostgreSQL · HNSW indexing · Semantic similarity search
Week 4: Production RAG (Retrieval-Augmented Gen)
Chunking strategies · Hybrid search (BM25 + Dense) · Re-ranking with Cohere · Hallucination mitigation
Week 5: Autonomous Agents & Tool Calling
ReAct agent loop · Function calling APIs · Multi-agent coordination · Stateful memory stores
Week 6: AI Logic Evaluation & Guardrails
Ragas & DeepEval metrics · Content moderation guards · Latency & cost optimization · LangSmith tracing
Week 7: Fine-Tuning & Custom Adapters
LoRA & QLoRA concept · Dataset curation & cleaning · Evaluation against base model · Edge deployment
Week 8: Autonomous AI Researcher Capstone
Multi-document synthesis agent · Live human-in-the-loop web UI · Presentation & Certificate Defense
How do I earn the AI Engineering & LLM Systems Sprint certificate?
Complete every weekly task and receive a reviewed or approved result of at least 65% on each. You can then purchase the optional ProofCohort certificate bundle. Partner-issued credentials require a confirmed, authorized issuing arrangement.
Read the assessment guide →Certificate verification and support →Prepare your project evidence
For each assigned task, connect its requirements to a deliverable the reviewer can open. Explain your contribution, decisions, checks, and any unfinished work. Use the actual task instructions for submission requirements and the templates below to organize your evidence.