Jędrzej Jamnicki Email & Phone Number

Poland

Wrocław, Pl

Development of chatbot used by all of the employees from the organization in day-to-day work. Agentic, multi-sourced RAG implemented in Azure environment with Typescript and Python. Utilizing OpenAI functions with LLamaIndex and Langchain. Publishing chat as web app in Teams environment.Development of LLM based context aware translation service with human in the loop (HITL) validation.Development of Azure Function, Container Apps in Python. Designing APIs in OpenAPI (swagger) specification, docs publishing and writing tests in Postman.

Wrocław, Dolnośląskie, Poland

Development of chatbot based on LLM (OpenAI API) and vector databases (Weaviate) with RASA and LangChain. I was responsible for data pipelines, prototyping of RAG (Retrieval-Augmented Generation) related solutions, and conducting preliminary analysis of data and results.Implementation of document level information extraction tools. I was responsible for combining and evaluating the NER and NEL models developed in PyTorch, transformers, spaCy, by team members, and for developing RegEx patterns. Participating in the 2+1 style annotation and validation with clients.Implementation of document processing services ex. chunking. Co-development of asynchronous client for Kubernetes data processing services. Creation of an application with Flask, Jinja, and JavaScript for validating XML files.During the trial month, I prototyped an authorization and authentication system based on OpenID Connect protocol (managed with Keycloak). The application has been developed with Flask, Jinja templates, JavaScript, Docker, and MySQL.Ad hoc: data manipulation, prototyping, and data quality assurance (DQA) related tasksTech: Docker, SQL, Python[pandas, spaCy, PyTorch, transformers, LangChain, Weaviate, ...]

Master'S Degree, Artificial Intelligence, 5.0 (Very Good)

[EN] Thesis on: Semi-Automated Cost-Efficient Large Language Model Data Preparation Techniques Daily experiments with models in PyTorch.

Engineer'S Degree, Systems Engineering, 5.0 (Very Good)

[PL] Thesis on: System for Training-based Expansion of Tools for Proper Name Mentions Recognition Based on Active Learning, Course.