Building and Deploying LLM Models from Scratch Using Only PyTorch
Date: 16.02.2026
Speaker: Dilyan Grigorov – Software Developer passionate about Python and Deep Learning, Stanford AI Master's student
About the event:
This webinar is part of the Python user group.
Webinar program:
The webinar “Building and Deploying LLM Models from Scratch Using Only PyTorch” is a practical session aimed at developers and ML engineers who want to deeply understand how large language models are built – without high-level ready libraries and without “black boxes.” We will cover the entire process – from architecture and training to optimization and production deployment.
In the webinar you will learn:
- How the architecture of LLM models works on a low level;
- How to implement Transformer blocks using only PyTorch;
- How to prepare and tokenize text data for training;
- How to build a training pipeline from scratch;
- Techniques for stable and efficient large model training;
- Memory and performance optimization during training;
- Fine-tuning and model evaluation;
- How to prepare the model for inference;
- Practical approaches to deployment as an API service;
- Best practices for scaling and monitoring.
Suitable for:
- ML engineers
- Python/PyTorch developers
- Data scientists
- AI enthusiasts with basic knowledge of deep learning
The webinar focuses on real examples, code, and architectural solutions directly applicable to practical projects.
Q&A session
The event is free but requires prior registration.
Python User Group
This event is part of the Python user group. Every month we organize Python-related events. To stay updated, subscribe to the group.
Organizer:
DEV.BG is a specialized job board for IT announcements.
Python Job Listings: If you are looking for a job related to Python technology, check the listings here.
When: 16.02.2026 at 19:30
We are looking forward to seeing you!
About the Speaker
Dilyan Grigorov is a software developer with a passion for Python and Deep Learning. He is an advocate for open source and the Python language itself. With 16 years of experience in Python programming, he has spent 5 years applying AI solutions. He is part of the organization team of one of the largest Python events – EuroPython.
Dilyan is a Stanford Master's student in Artificial Intelligence, studying under mentors such as Andrew Ng, Fei-Fei Li, and Christopher Manning. His mentors include software engineers and AI experts from Google and Nvidia.
His passion for AI and ML stems from his experience as an SEO specialist working with search engine algorithms daily. He enjoys engaging with the software community by frequently presenting at local meetups and larger conferences. In his free time, he enjoys reading, hiking, long walks, and playing piano.
Event sponsor:
Man Group
Ticket Options
FREE
Free event
All prices shown are relevant to the date when the event was added to the website. Current prices may differ.