
Przemysław Przybyszewski
Verified Expert in Engineering
AI and Software Developer
Warsaw, Poland
Toptal member since September 1, 2020
Przemysław has a Ph.D. in economics and a Master's degree in data science. He enjoys developing AI projects and has an exceptional understanding of how to use data to generate profitable solutions to some of the industries toughest problems. He co-authored a research paper presented at the MICCAI conference (the premier international conference in information processing, machine learning, and computational modeling) and developed an anti-fraud user behavior anomaly detection algorithm.
Portfolio
Experience
- Artificial Intelligence (AI) - 5 years
- Machine Learning - 5 years
- Statistics - 4 years
- Deep Learning - 3 years
- Data Engineering - 3 years
- Natural Language Processing (NLP) - 2 years
- Speech Recognition - 2 years
- Speech-to-Text (STT) - 1 year
Preferred Environment
Windows, Vim Text Editor, Visual Studio Code (VS Code), MacOS, IntelliJ IDEA, PyCharm, Linux
The most amazing...
...architecture I designed with a team of developers that I led was for a software product that enabled auditing and a flow of data science projects.
Work Experience
AI/Data Architect and Engineer | Data Scientist
Self employment
- Developed an LLMOps pipeline and prepared multiple agentic workflows (models both trained in-house and 3rd-party provided), speeding up the software development cycle (AI-driven code reviews, initial PR for the tickets, code template generation).
- Developed AI agentic workflows helping out the GTM team in their daily routine tasks (RFC evaluation and draft preparation, internal-resource agentic chatbot, opportunity price evaluator).
- Provided an LLM agentic pipeline for extracting invoice data from an invoice page in any format. Achieved around 92% accuracy on all invoice fields (monthly volume of 1.5 million invoices).
Software Developer
Cherrypick Games
- Developed a deep learning model to analyze the sequence of in-game events to predict the chances of a given user being a potential spender in the game.
- Prepared the architecture and implementation of the entire data infrastructure, including a data-lake from different sources through DataFlow in BigQuery, scheduling queries for data management, and a BI board for administration.
- Designed multiple ad-hock queries to support management's strategic decisions regarding mobile game development.
AI Engineer
Stampli
- Worked on a real-time pipeline enabling the extraction of information from an invoice (OCRing the incoming files and running agentic AI workflows/LLMs/RAGs containing historical data) to ensure the information extracted from the invoice is correct.
- Built the initial system design for AI-assisted proposal generation for procurement requests (automatic approver suggestion, product proposal).
- Contributed to the deployment and enhancement of a Python FastAPI service that can handle the load of real-time processing (various dimensions) of tens of thousands of invoices per day.
Data Architect/Engineer
BJ's Wholesale Club - Marketing/Analytics
- Designed and implemented the initial version of a refactor MLOps pipeline, enabling model deployment, retraining, and inference triggered by code changes or data events.
- Managed a roadmap for development features in acquisition and personalization engines, overseeing bug fixes, feature ticket preparation, PR reviews, and ensuring compliance with functional and non-functional requirements.
- Led a team of 4 developers, providing guidance, removing blockers, and reviewing their work to ensure high-quality implementation.
- Optimized ETL processes by migrating to AWS Glue and EMR serverless for cost efficiency and fine-tuning Spark jobs based on execution plans to enhance performance.
Software Architect | Back-end Engineer
Lumilook
- Prepared a GenAI tool, which generated safety recommendations for the company based on the statistics of occurring events, their location, and data gathered from safety managers through AI-assisted conversation and past safety reports.
- Designed and implemented the back end for processing streaming data, analyzing them in tumbling windows to generate real-time alerts for safety managers. Also prepared a device capable of collecting data from CCTV cameras and running AI inference.
- Prepared an API service to visualize safety incidents across different places in the warehouse at different times.
Security Software Engineer
ByteDance AI Lab
- Developed an anti-fraud user behavior anomaly detection algorithm by which we could effectively block IPs used by bots.
- Prepared a POC of an AI-powered WAF and IDS in the company's internal cloud environment.
- Explored the usage of eBPFs in enabling real-time network traffic analysis with the use of machine learning models from the user space.
Senior Software Developer
deepsense.ai..
- Implemented multiple features in Scala/Java in a Kubernetes environment for a Neptune project, a machine learning experiment management tool.
- Implemented and maintained multiple microservices (Go, Python, Java) for a product; a one-click deployment script for preparing a cloud-agnostic environment (worked on GCP, AWS, and Azure) for data scientists.
- Assisted with an AI pipeline for generating a list of ingredients from the images of FMCG products (extracting ROI on images through Fast and FasterRCNN, running OCR, and then applying FastText on the returned content to get the ingredients list).
Member of the Research Team
Interdisciplinary Centre for Mathematical and Computational Modelling
- Prepared a multimodal deep learning model for estimating the healing progress of the Achilles tendon based on the sequence of US and MRI scans.
- Prepared two microservices (Java) for the data management of model training and experiment tracking.
- Co-authored a research paper presented at the MICCAI conference. (The premier international conference in information processing, machine learning, and computational modeling in medical image computing and computer assisted interventions).
Experience
Context Cartographer
Data Architect/Engineer
Chatbot for Serving Loans for Construction Developers
Education
Ph.D. in Economics
Warsaw School of Economics - Warsaw, Poland
Master's Degree in Data Science
Warsaw School of Economics - Warsaw, Poland
Bachelor's Degree in Computer Science
University of Warsaw - Warsaw, Poland
Bachelor's Degree in Quantitative Methods in Eonomics
Warsaw School of Economics - Warsaw, Poland
Certifications
Google Cloud Certified Professional Data Engineering
Skills
Libraries/APIs
TensorFlow, Keras, PyTorch, Pandas, Scikit-learn, NumPy
Tools
Amazon Athena, PyCharm, IntelliJ IDEA, Vim Text Editor, Flink, ChatGPT, AWS Step Functions, Apache Airflow, GitLab CI/CD, Amazon Elastic MapReduce (EMR), Amazon SageMaker, Google Cloud Dataproc, Cloud Dataflow, Google AI Platform, Amazon OpenSearch
Languages
Python, Java, Go, Scala, SQL, R
Paradigms
Automation, DevOps, Concurrent Programming
Frameworks
Spark
Platforms
Google Cloud Platform (GCP), Kubernetes, Linux, Visual Studio Code (VS Code), Amazon Web Services (AWS), AWS Lambda, AWS IoT, AWS IoT Greengrass
Storage
ClickHouse, Amazon DynamoDB, Google Cloud SQL
Other
Artificial Intelligence (AI), Machine Learning Operations (MLOps), Generative Pre-trained Transformers (GPT), Forecasting, API Integration, Data Analytics, Data Engineering, Machine Learning, Data Science, Statistics, Fine-tuning, Natural Language Processing (NLP), Speech Recognition, Convolutional Neural Networks (CNNs), Neural Networks, Predictive Modeling, AI Design, Google BigQuery, Big Data, Data Analysis, Finance, Monte Carlo Simulations, Financial Modeling, RAG Systems, Large Language Models (LLMs), Large Language Model Operations (LLMOps), Bayesian Inference & Modeling, Bayesian Statistics, Deep Learning, Reinforcement Learning, Deep Reinforcement Learning, Stable Diffusion, LoRa, Speech-to-Text (STT), Text-to-Speech (TTS), OpenAI, Generative Artificial Intelligence (GenAI), Amazon Kinesis, Amazon Timestream, Google Cloud ML, Software Development, Vector Search
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
Share your needs
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring