Working on application of large language models

I delivered 8 models in 6 months, there are a couple of them:- Community detection based on the textual features. The model is able to detect the main authors' communities (like religion, and hobbies). I built it using: Neo4j, Kafka.- Gender prediction of a blogger audience, based on Mixture Density Network models. I implemented it using PyTorch and optimized network structure using Optuna.- Topic modeling with a stable transition between runs. I used Python3.8, Network-x, and… Show more I delivered 8 models in 6 months, there are a couple of them:- Community detection based on the textual features. The model is able to detect the main authors' communities (like religion, and hobbies). I built it using: Neo4j, Kafka.- Gender prediction of a blogger audience, based on Mixture Density Network models. I implemented it using PyTorch and optimized network structure using Optuna.- Topic modeling with a stable transition between runs. I used Python3.8, Network-x, and Hugging Face transformers.All models were served as REST services using MLFlow.Last 5 months I was working on a research project. Developed a few different models that identify big5 traits based on textual features. Show less

Optimized Monte Carlo sampling module using profiling. The speed up was on average by 2 times. Participated in an internal conference with a well clustering algorithm (optimization of the choice of the position of the well and their combination into clusters to reduce the infrastructure cost).