AI Engineering: The Developer's Path

Master the fundamental tool that every professional developer uses daily. Learn to track changes, collaborate with others, and manage your code like a pro from the very beginning of your development journey.

Python serves as the foundation of modern data science, providing essential programming skills for data manipulation, analysis, and machine learning. This module develops your Python proficiency from basics through data science applications.

This module introduces students to fundamental database design principles and SQL (Structured Query Language). Students will learn how to design efficient relational database schemas, implement entity-relationship diagrams, normalize databases, and write SQL queries to create, retrieve, update, and delete data. The course covers both theoretical concepts and practical applications, with hands-on exercises using industry-standard database management systems. By the end of this module, students will be able to design and implement a functional database solution for real-world applications.

Build production-grade REST APIs with FastAPI and Pydantic — authentication, async endpoints, database integration, caching, testing, and auto-generated documentation. The backbone every AI service exposes to the outside world.

This foundational module covers traditional machine learning before diving into deep learning. You'll master the complete ML workflow using scikit-learn: data preparation, model building, evaluation, and improvement. Topics include supervised vs unsupervised learning, regression models (Linear, Polynomial), classification models (Logistic Regression, Decision Trees, Random Forests, SVM), clustering (K-Means), and essential evaluation metrics (Accuracy, Precision, Recall, F1, ROC-AUC). By understanding classical ML first, you'll appreciate why and when deep learning is needed.

Work with state-of-the-art language models.

This module covers LLM architecture internals (GPT, LLaMA), efficient fine-tuning techniques (LoRA, QLoRA), Retrieval-Augmented Generation for production applications, and a comprehensive capstone project to demonstrate your skills.

Move from "asking an LLM nicely" to a discipline — structured prompts, chain-of-thought, output formatting, prompt injection defence, testing, and versioning. The cheapest, most underrated lever in any AI system.

Stand on the shoulders of LangChain, LlamaIndex, Haystack, and MCP. Patterns for composing models, retrievers, memory, and tools without reinventing the plumbing every project.

Give an LLM access to your data without fine-tuning. Embedding models, vector databases, chunking strategies, hybrid search, re-ranking, advanced patterns (parent-document, RAPTOR, Graph RAG), evaluation, and production deployment.

From single-tool ReAct loops to multi-agent crews. LangGraph state machines, tool design, memory, safety guardrails, human-in-the-loop checkpoints, and the observability you need to debug an agent in production.

The product layer most AI tutorials skip. Streaming, latency, AI-specific error states, conversation UI, feedback loops, trust indicators, accessibility, and multimodal input — what makes an AI feature feel finished.

Build evals from scratch, write golden datasets, run LLM-as-judge pipelines, red-team your own system, and ship regression tests so a prompt or model swap can't silently break things. The discipline that separates demos from products.

When an LLM, agent, or RAG pipeline misbehaves, where do you look? Tracing reasoning failures, retrieval-vs-generation triage, token/context audits, embedding-quality analysis, and the LangSmith/Langfuse workflows that turn a black box into a debuggable system.

Token economics, prompt compression, semantic caching, model routing, batching, and the build-vs-buy spreadsheet that decides whether an AI feature is viable. Often the difference between a profitable product and a science project.

Containerise, deploy, observe, and secure AI services in the wild. Docker patterns for inference, cloud deployment on AWS / GCP / Azure, Kubernetes basics, CI/CD via GitHub Actions, observability with LangSmith / Langfuse, and the API-key + prompt-injection guardrails that keep production safe.

Fine-tuning, multimodal, compound AI systems, small language models, alignment, ethics — plus your portfolio-grade capstone: an end-to-end AI application with evaluations and production deployment.