Михаил Гончаров - CV Инженер Машинного Обучения

Data analysis closely supported Natural Language Processing Engineering. • LSEG SDK Co-pilot project: performed evaluations of code, generated by Large Language Models (LLM). Identified error origins and bottlenecks and delivered action items. • Engineered an automated evaluation pipeline integrated with unit tests, detecting over 94% of errors for different LLMs. Contributed quality and error analysis within complex prompt engineering pipelines. • Financial Research Reports Similarity Tool: developed a tool that gets the top n documents with the highest similarity to the input provided. Utilizing the OCR model (PDFTron) and semantic search to identify similar financial reports, facilitating comparison of annotators' labels and optimizing inter-annotator agreement through a Streamlit UI. • Provided a comprehensive analysis of text classification, assessing its effectiveness and suggesting enhancements for superior accuracy. Automated it with a clean and readable Python script.

BSc., MSc., Moscow Auto And Road Construction University (MADI) Mechanical Engineer

Large Language Models (LLM) IR (RAG, semantic search, tf-idf) Prompt Engineering Classical ML, Data Analysis Streamlit, Unit Testing, SQL

Английский

I am an ML Engineer by training with 4 years of practical experience in data analysis and development during which I wrote production code and worked on full DS project cycles. I am enthusiastic about researching diverse solution techniques.
Currently, I am looking for positions as a Data Scientist, ML/NLP Engineer or Data Analyst within a data science team.