Building expertise in Large Language Models through hands-on projects, starting from basics and advancing to cutting-edge implementations
📖 Large Language Model (LLM) Projects – Roadmap & Progress
Large Language Models (LLMs) represent the cutting edge of AI, and I’m actively building a portfolio around how they can be applied to finance, data science, and decision support.
This section is not a finished showcase yet — it’s my plan and roadmap. As I make progress in LLM and Generative AI projects, I will continue updating this page with hands-on builds, experiments, and learnings.
🎯 My Focus Areas
- 🤖 AI Assistants — prototypes that support financial analysis and reporting
- 📊 Automated Insights — summarisation and exploration of large, complex datasets
- 📚 Knowledge Management Tools — retrieval-augmented systems for finance and analytics
- 🧪 Prompt Engineering & Fine-tuning — domain-specific experiments for finance/data science
- ⚙️ Integrations & Deployment — APIs, Streamlit apps, Docker-based LLM environments
Each project will highlight both the technical experimentation (working with GPT, LLaMA, Ollama, etc.) and the strategic business value — showing how LLMs enable faster insights, automation, and smarter decisions.
🪜 The AI Ladder – My Learning Path
I approach mastery as a ladder, climbing one step at a time. This roadmap reflects where I am now and what I’m working on next.
Step 1 – Mathematical Foundations (✅ In Progress)
- Building strong foundations in linear algebra, probability, statistics, calculus.
- Using 3Blue1Brown for visual understanding and The Palindrome for structured explanations.
Step 2 – Hands-on Practice (🔄 Ongoing)
- Applying theory using Scikit-Learn and PyTorch.
- Following Sebastian Raschka’s ML with PyTorch & Scikit-Learn to strengthen model building skills.
Step 3 – Data Engineering (📌 Next Focus)
- Developing comfort with data pipelines, storage, cleaning, and transformation.
- Currently studying Designing Data-Intensive Applications to master scalability and reliability.
Step 4 – ML / AI Engineering (📌 Upcoming)
- Moving from experiments to robust ML/LLM pipelines.
- Learning MLOps/LLMOps, experiment tracking, and deployment.
- Resources: Designing Machine Learning Systems & AI Engineering (Chip Huyen).
Step 5 – Build, Build, Build (🚀 Ongoing)
- Applying learning to real-world projects in finance, analytics, and decision support.
- Workflow: Pick → Build → Break → Improve → Repeat.
- Portfolio updates will showcase progress step by step.
🔭 Next Steps
- 📌 Begin publishing LLM-powered finance/data projects on GitHub.
- 📌 Share experiments in prompt engineering and fine-tuning.
- 📌 Build Streamlit + API apps powered by LLMs.
- 📌 Explore retrieval-augmented generation (RAG) for financial knowledge bases.
- 📌 Document progress regularly to show growth along the AI Ladder.
This is a living roadmap. As I climb further in my LLM and Generative AI journey, I’ll keep updating this space with projects, insights, and lessons learned.
