Explore our comprehensive catalog of micro-credential courses designed to boost your skills and career. Flexible, industry-aligned, and recognized by employers.
⏱ Duration: 24 Sessions | 36 Contact Hours (1.5 Hours per Session) | Mode: Physical / Hybrid
Mahbuba Yesmin Turaba – Founder and CEO, MindSpark; Adjunct Faculty, Canadian University of Bangladesh (Industry Expert / Partner)
Abdus Sattar – Associate Professor, Department of CSE, Daffodil International University
Taslima Akhter – Lecturer, Department of CSE, Daffodil International University
This course is designed to meet growing industry demand by providing learners with practical knowledge of modern AI technologies, including Google AI Studio, Transformers, Retrieval-Augmented Generation (RAG), FastAPI, and vector databases. It guides participants through the complete end-to-end journey—from identifying a viable business idea to developing, containerizing, and deploying a working AI Minimum Viable Product (MVP) on modern cloud infrastructure, bridging the critical gap between foundational AI theory and production-ready application development.
Industries are rapidly moving beyond simple AI tools and wrapper concepts, seeking professionals who can build practical, domain-specific AI applications for real-world problems.
Although many graduates and early-career developers understand basic AI concepts, job-market analysis and industry reports indicate a critical shortage of professionals equipped to connect Large Language Models (LLMs) with private company or domain-specific datasets. Organizations now demand engineers who can build end-to-end data pipelines, implement vector search, prevent hallucinations, and deploy working applications into production. This micro-credential course directly responds to that market gap through hands-on, project-driven training.
Purpose: To master the end-to-end pipeline required to convert an AI concept into a deployment-ready MVP using Google AI Studio, advanced RAG architectures, Vector Databases, FastAPI, and modern cloud infrastructure.
This course aims to:
Identify AI Business Opportunities: Distinguish between generic AI wrapper ideas and viable AI products, mapping out real-world opportunities across education, healthcare, e-commerce, banking, and agriculture.
Master User-Centric Design Frameworks: Apply Design Thinking methodologies, Jobs-to-Be-Done (JTBD) frameworks, and Value Proposition Canvases to transform raw concepts into validated problem statements.
Source & Extract Unstructured Data: Gain hands-on experience collecting, web scraping, and parsing data from heterogeneous sources (PDFs, web pages, CSV, Excel files) while maintaining ethical standards and data privacy.
Implement Data Cleaning & Chunking Pipelines: Build automated processing pipelines using Python to eliminate noise, strip irrelevant content, normalize whitespace, and implement semantic/hierarchical chunking strategies.
Master LLM Foundations & Google AI Studio: Develop a solid functional understanding of Transformer architectures (attention mechanisms, tokenization, context windows, decoding parameters) and rapidly prototype using Gemini models, system instructions, and structured prompting in Google AI Studio.
Build Advanced Retrieval-Augmented Generation (RAG) Systems: Construct robust end-to-end RAG pipelines, progressing from baseline setups to advanced techniques including multi-stage retrieval, metadata filtering, reranking, and dynamic context injection.
Ship & Showcase an End-to-End MVP: Package pipelines into FastAPI backend microservices, apply safety guardrails, deploy to cloud infrastructure, and present a functional, domain-specific AI MVP accompanied by a future production roadmap to stakeholders.
(Delivered across 12 Core Modules / 24 Sessions, 1.5 Hours per Session — Total 36 Contact Hours)
Module 1 (Sessions 1–2): AI Product Discovery & Value Proposition
Module 2 (Sessions 3–4): Transformer Foundations & LLM Fundamentals
Module 3 (Sessions 5–6): AI-Assisted Rapid Prototyping with Google AI Studio
Module 4 (Sessions 7–8): Unstructured Data Engineering & Advanced Chunking
Module 5 (Sessions 9–10): Embeddings, Vector Spaces & Vector DBs (ChromaDB / Pinecone)
Module 6 (Sessions 11–12): Retrieval-Augmented Generation (RAG) Core & Quality Optimization
Module 7 (Sessions 13–14): Backend API Development (FastAPI) & System Integration
Module 8 (Sessions 15–16): RAG vs. Fine-Tuning Decision Framework & Dataset Preparation
Module 9 (Sessions 17–18): LLM Evaluation Frameworks & Quality Metrics
Module 10 (Sessions 19–20): AI Safety, Alignment, Guardrails & Collaborative Git Workflow
Module 11 (Sessions 21–22): Deployment, Cloud Infrastructure (Google Cloud Run / Hugging Face Spaces) & LLMOps Preview
Module 12 (Sessions 23–24): Final Project Showcase & Production AI Roadmap
The course is heavily weighted toward hands-on implementation, lab simulation, and end-to-end product engineering (accounting for 90% of total assessment weight via practical labs and functional MVP development):
Hands-On Workshops & Lab Simulations: Interactive lab sessions covering ethical web scraping, PDF/tabular data extraction, recursive and semantic chunking pipelines, and vector indexing in ChromaDB and Pinecone.
Rapid Prototyping & Pipeline Engineering: Developing real-time prompts in Google AI Studio, implementing query rewriting and reranking in Python, and wrapping RAG workflows into production FastAPI microservices.
End-to-End MVP Capstone Build & Cloud Deployment: Learners identify a domain-specific problem across industries (healthcare, finance, education, e-commerce, or agriculture), build a complete AI application with input/output safety guardrails and hallucination checks, deploy it via Git and cloud platforms, and pitch the live system architecture to an evaluation panel.
After completing the course, participants will be able to:
LO1: Identify real-world domain problems across industries (healthcare, finance, education, e-commerce, agriculture) and apply Design Thinking, Jobs-to-Be-Done (JTBD), and Lean Canvas frameworks to evaluate AI product-market fit before writing code.
LO2: Collect, scrape, and clean unstructured datasets (PDFs, web text, tabular files) under ethical guidelines, applying advanced chunking strategies (recursive, layout-aware, semantic) and metadata enrichment to prepare high-quality vector index inputs.
LO3: Generate text embeddings using modern models, store and index vector representations in databases (ChromaDB/Pinecone), and write vector queries using distance metrics and metadata filters to retrieve domain context accurately.
LO4: Build and optimize Retrieval-Augmented Generation (RAG) pipelines in Python incorporating query rewriting and reranking, combined with dynamic prompt injection into Gemini models via Google AI Studio APIs to synthesize grounded, context-aware answers.
LO5: Package RAG pipelines into production-oriented FastAPI web applications, implementing input/output safety guardrails, hallucination checks, rate limiting, and automated evaluation metrics (precision, recall, faithfulness).
LO6: Containerize and deploy AI backend applications securely using Git workflows and cloud infrastructure (Google Cloud Run / Hugging Face Spaces), presenting a live demonstration detailing system architecture, evaluation metrics, and a future production roadmap.
Undergraduate and Graduate Students (CSE, Software Engineering, Data Science, or related disciplines)
Early-Career AI Engineers and Software Developers
Data Analysts looking to transition into Generative AI engineering
Tech Entrepreneurs aiming to build AI-powered products and startups
Enrolled in or graduated from an Undergraduate/Graduate program in Computer Science, Engineering, or a related analytical field, OR working as an early-career developer/analyst
Foundational knowledge of Python programming (required for data pipelines, RAG implementation, and FastAPI backend integration)
Basic familiarity with web technologies, APIs, and Git version control is advantageous
After completing this micro-credential, learners may:
Enter Industry Roles: Work as an AI Engineer, Generative AI Developer, or AI Application Developer.
Launch Ventures: Pursue entrepreneurial opportunities by building and commercializing AI-powered products and SaaS solutions.
Progress Academically & Professionally: Advance toward Higher Study, Advanced Micro-Credentials, or specialized Professional Certifications in LLMOps and Enterprise AI Architecture.
Well-equipped multimedia classrooms
Dedicated computer labs (Available at DIU) with internet connectivity
AI & LLM Platforms: Google AI Studio, Gemini APIs, Transformer Models
Programming & Backend: Python, FastAPI
Vector Databases: ChromaDB, Pinecone
Deployment & DevOps: Git, Docker/Containerization, Google Cloud Run, Hugging Face Spaces
Product Design Frameworks: Design Thinking, Jobs-to-Be-Done (JTBD), Value Proposition Canvas, Lean Canvas
Course syllabus and structured session slides
Python codebooks and data engineering/RAG templates
Product discovery canvases (JTBD, Lean Canvas, Value Proposition templates)
LLM evaluation and safety guardrail checklists
Fully mapped to an Outcome-Based Education (OBE) scheme using Projects, Practicals, and Presentations:
Hands-On Lab Work – 10% (Assesses LO1–LO3: Product discovery, data scraping/chunking, and vector database indexing)
Functional Product Development Project – 80% (Assesses LO4–LO6: End-to-end RAG pipeline, FastAPI integration, guardrails, and cloud MVP build)
Report & Group Presentation – 10% (Assesses All LOs: Includes live demonstration of system architecture, evaluation metrics, and production roadmap)
Upon successful completion of all assessments, participants will be awarded an official Micro-Credential Certificate from the Micro-Credential Academy, Daffodil International University.
DIU Students & Alumni: BDT 4,000
External Candidates & Entrepreneurs: BDT 4,500
Join thousands of students already enrolled
Try adjusting your filters to see more courses.