Dmitrii Matveev Email & Phone Number

Moldova

Chișinău, Moldova

Responsibilities:• Handling outsourcing tasks for a big EU bank, which include but are not limited to:• Developing and implementing ChatBot applications and AI Pipelines powered by OpenAI API to handle hundreds of queries per week from internal employees.• Integrating AI functionalities into backend applications using FastAPI.• Collecting datasets for further model fine-tuning.

Achievements & Responsibilities:• Released updates to implement scaling and adaption of a product for new warehouse customer.• Successfully migrated computer vision models from PyTorch and TensorFlow to ONNX, assisted in ONNX integration into C++ code.• Ported product inference to a Docker container.• Assisted in project development managing for a team of 2 programmers.Projects: real-time video control system for warehouse working with barcode/datamatrix/QRcode recognition.

Achievements:• Successfully released MVP for marketplace goods matching with 70% quality in a short timeframe, improved the result close to 90% quality with a sentence transformer neural network.• Achieved 90% speed-up of production on the same server machine by processing logic changing and multiprocessing applying.• Implemented advanced text data preprocessing for parsing e-commerce goods' attributes (libraries used: pandas, regex, spaCy).• Improved Argo/Kubernetes manifests and documentation.Responsibilities:• Participate in developing solutions for top brands.• Collaborate on business ideas and decompose them into technical tasks.• Train and release Artificial Intelligence tools for e-commerce analytics team.• Containerise projects with Docker and set up Argo pipeline.• Write clean and structured code.Projects: matching of marketplaces' goods with client's product catalog; grouping of similar marketplaces' goods; tool for comparing desired and factual SKUs name/description/images.

Achievements:• Released 2 Computer Vision B2B projects, where I trained key NN-models and developed core business logics.• Successfully checked business ideas in a short timeframe by creating MVPs and providing deep working reports.• Effectively assisted in project development management for both a data labelling team and a team of 2 programmers.• Participated in a tech exhibition on behalf of the company.Responsibilities:• Participate in the full cycle of product development.• Fine-tune and train Neural Network models: detectors, classifiers, super-resolution.• Create tasks for data labelling team and prepare datasets.• Analyse open-source projects, State-Of-The-Art (SOTA) Neural Network models, modern tools and approaches.Projects: real-time video control system for warehouse working with barcode/datamatrix/QRcode recognition; video analytics for counting buyers flow in a grocery store; car flow tracking and classification; analytics of queue duration of buyers.