Paweł is available for hire

Paweł Balon

Verified Expert in Engineering

Machine Learning Developer

Location

Warsaw, Poland

Toptal Member Since

February 22, 2022

Paweł is a former competitive programmer, and by the age of 25, he had already become the global head of AI research for UBS IB Tech. As a cross-functional technology leader, he's worked with teams across Poland, Singapore, the US, and the UK. Even as an executive, Paweł stayed close to tech, thoughtfully translating the newest AI research into business products. Paweł is known for designing one of the best information retrieval platforms in the world. He's fluent in Python and in love with NLP.

Algorithms Natural Language Processing (NLP)Machine Learning Computer Vision Time Series Analysis Python TensorFlow Recommendation Systems Deep Learning Artificial Intelligence (AI)Deep Neural Networks Predictive Analytics Software Development Clustering A/B Testing Rasa Competitive Programming RNN

Portfolio

UBS

Deep Learning, Natural Language Processing (NLP)...

Prile

JavaScript, WebAssembly, C++, Java, Algorithms, Cryptocoin Mining, Analytics...

Comarch

Deep Learning, Natural Language Processing (NLP)...

Experience

Algorithms - 13 years Competitive Programming - 11 years Python - 8 years Time Series Analysis - 7 years Natural Language Processing (NLP) - 7 years Machine Learning - 7 years Computer Vision - 7 years Finance - 7 years

Availability

Part-time

Preferred Environment

Python, Linux RHEL/CentOS, Git

The most amazing...

...product I've designed is the information retrieval platform. With complex NLP hidden behind a simple GUI, it was quickly adopted among dozens of UBS' divisions.

Work Experience

Global Head of AI Research for Investment Banking Technology

2018 - 2021

UBS

Led AI research globally. Set up directions for research, connected multiple business cases with emerging technological opportunities, consulted and coordinated cross-divisional projects, and assessed in-house development versus external solutions.
Focused on NLP projects to reduce operations costs (by automating repetitive, time-consuming processes) and allow people to make better, data-driven decisions (by applying AI to unleash potential and leverage the business value of UBS’ data).
Led core AI teams responsible for developing critical AI projects and platforms. Brainstormed at a high level (platform's use-cases scope, fundings, and more) and a low level (technical ideas and best way to implement them.).
Built from scratch global AI excellence center for UBS IB Tech based in Poland, participated in shaping standards for production deployment of AI solutions for UBS globally.
Formed and coordinated a global research community. Ran multiple upskilling initiatives and events, connected AI scientists among different UBS divisions, supported career switch into the Machine Learning area.
Evangelized AI to every kind of audience outside (major hackathons, top universities, conferences) and inside the bank which included global learning platform and senior management (GEB members) .

Technologies: Deep Learning, Natural Language Processing (NLP), GPT, Generative Pre-trained Transformers (GPT), Computer Vision, Python 3, Language Models, TensorFlow, PyTorch, Algorithms, Finance, Stock Market, Predictive Modeling, Information Retrieval, Analytics, OpenCV, Machine Learning, Python, Research, Keras, Data Science, Deep Neural Networks, Computer Vision Algorithms, Convolutional Neural Networks (CNN), Neural Networks, Financial Data, Financial Modeling, Automation, Scripting, Image Processing, Models, OCR, Data Visualization, Big Data Architecture, Big Data, Back-end, Datasets, Data Pipelines, Public Speaking, AI Programming, Chatbots, Minimum Viable Product (MVP), Proof of Concept (POC), Education, Pandas, Natural Language Understanding (NLU), Matplotlib, Contract, Docker, GitHub, Hugging Face, PyTorch Lightning, Reinforcement Learning, Financial Forecasting, Financial Engineering, SQL, Software Architecture, Technical Leadership, Full-stack, Apache Kafka, Kubernetes, Rasa NLU, CTO, Leadership, REST APIs, Text Recognition, Software as a Service (SaaS), Databases, Hedge Funds, Recurrent Neural Networks (RNNs), Explainable Artificial Intelligence (XAI), Text to Speech (TTS), Workshop Facilitation, Data-driven Decision-making, Decision Modeling

Co-founder

2017 - 2018

Prile

Co-founded a startup to move in-browser mining from being used mainly through crackers to legit products for companies/individuals, thus creating a new way to monetize internet content.
Co-developed the world’s fastest JavaScript and WebAsm script mining for Monero cryptocurrency in the browser.
Built fully scalable server architecture capable of managing millions of requests per minute deployed in AWS cloud and designed a secure Monero wallets management system based on the hot/cold wallet idea.

Technologies: JavaScript, WebAssembly, C++, Java, Algorithms, Cryptocoin Mining, Analytics, Research, Amazon Web Services (AWS), Scripting, Back-end, Blockchain, Startups, Early-stage Startups, Public Speaking, Minimum Viable Product (MVP), Proof of Concept (POC), Docker, GitHub, Financial Engineering, SQL, Software Architecture, Technical Leadership, Full-stack, Kubernetes, CTO, Leadership, REST APIs, Software as a Service (SaaS), Databases

Head of AI Research for Financial Services R&D

2016 - 2018

Comarch

Led AI research in the R&D department for financial services which included building multiple AI POC solutions for companies in the financial, insurance, and banking domains as well as working with structured data, computer vision, and NLP.
Led the development of an anti-money laundering POC built for a major bank. By combining various approaches, we defined the new SOTA accuracy. Comarch recognized this opportunity and built a new department solely to grow the solution.
Designed and developed the NLP and algorithm of Myra: a conversational in-car investment assistant. Myra was presented at Finnovate 2017 in London and got a lot of interest from primarily Polish but also the international press.
Represented Comarch during multiple conferences, hackathons, and workshops and increased the recognition of the Comarch AI.
Conducted technical interviews and found promising candidates from university connections and my social network.

Technologies: Deep Learning, GPT, Generative Pre-trained Transformers (GPT), Natural Language Processing (NLP), Artificial Intelligence (AI), Voice Recognition, TensorFlow, Algorithms, Genetic Algorithms, OpenCV, Analytics, Computer Vision, Machine Learning, Python, Research, Keras, Data Science, Deep Neural Networks, Computer Vision Algorithms, Convolutional Neural Networks (CNN), Neural Networks, Financial Data, Financial Modeling, Automation, Scripting, Signal Processing, Digital Signal Processing, Image Processing, Models, Data Visualization, Video Processing, Video Analysis, Back-end, Datasets, Text Generation, Data Pipelines, Speech Recognition, Public Speaking, Voice Analysis, AI Programming, Chatbots, Minimum Viable Product (MVP), Proof of Concept (POC), Education, Pandas, Natural Language Understanding (NLU), Matplotlib, Contract, Docker, GitHub, Reinforcement Learning, Financial Forecasting, Financial Engineering, SQL, Software Architecture, Technical Leadership, Full-stack, CTO, Leadership, REST APIs, Text Recognition, Software as a Service (SaaS), Databases, Hedge Funds, Recurrent Neural Networks (RNNs), Explainable Artificial Intelligence (XAI), Text to Speech (TTS), Speech to Text, TortoiseGit, Workshop Facilitation, Data-driven Decision-making, Decision Modeling

Software Developer Intern in a High Performance Computing Team

2016 - 2016

Samsung R&D Institute

Recognized and trusted with conversion of a couple core OpenCV (image processing) algorithms to be DSP (new mobile processors) efficient including the whole process from design through C++ implementation till ASM-level optimizations.
Overtook final code reviews for interns and provided them with guidelines to create clean and efficient code.
Conducted a series of lectures about algorithms for competitive programming from his initiative.

Technologies: Computer Vision, C++, Algorithms, Assembler, DSP, OpenCV, Analytics, Machine Learning, Python, Research, Data Science, Computer Vision Algorithms, Scripting, Signal Processing, Digital Signal Processing, Image Processing, Models, Data Visualization, Datasets, AI Programming, GitHub

Experience

Information Retrieval Platform

One of the platforms I have designed while working at UBS. It automatically extracted and understood information even from poor quality scans of complex legal documents written in technical language. Under the hood, the platform leveraged the newest transformer research boosted by complex algorithms. In the front end, it was an intuitive web page easy to use even by non-technical people. The platform significantly outperformed market solutions and offered jointly processing of both graphical layout and the text. In the first year of development, it became an integral part of dozens of different business use-cases. For this platform, I was leading the core development team and have organized multiple satellite deployment teams across other UBS divisions, which were on-boarding user-cases from their business organizations to the platform.

ML for AML - Automation of Anti-money Laundering Regulatory Compliance

One of the PoCs I have built for a central bank while working at Comarch R&D was a solution based on the digital history of transactions, deciding whether the subject was involved in money laundering activities. By combining deep neural network classifiers to look for patterns that repeat over history and unsupervised anomaly detection to spot new ones early, it turned out to achieve superior accuracy (even while trained on a small amount of data). It was called the future of anti-money-laundering by stakeholders at the bank and recognized as an ample opportunity by Comarch, who decided to build a new department solely to grow the solution.

NLP for Stock Recommendations

This project was in cooperation with UBS wealth management and one of the last projects I kicked off before leaving UBS. There was already a recommendation system that helped UBS’ analysts offer stock to the clients according to their personalized preferences. We have extended that system with an NLP component that understands what’s happening on the markets based on financial news from multiple sources.

Skills

Libraries/APIs

PyTorch, TensorFlow, XGBoost, Pandas, Scikit-learn, NumPy, Matplotlib, OpenCV, Keras, PyTorch Lightning, Social Media APIs, Rasa NLU, REST APIs

Tools

ChatGPT, GitHub, Rasa.ai, TortoiseGit, Git

Languages

Python, Python 3, C, SQL, XML, C++, JavaScript, Java, Assembler

Paradigms

Anomaly Detection, REST, Data Science, Parallel Programming, Automation

Platforms

NVIDIA CUDA, Blockchain, Docker, Apache Kafka, Kubernetes, Linux RHEL/CentOS, Amazon Web Services (AWS)

Storage

Data Pipelines, Databases

Other

Natural Language Processing (NLP), Machine Learning, Computer Vision, Recommendation Systems, Time Series Analysis, Deep Learning, Optimization, Language Models, Algorithms, Stock Market, Predictive Modeling, Information Retrieval, Genetic Algorithms, Cryptocoin Mining, Information Extraction, Artificial Intelligence (AI), Natural Language Understanding (NLU), Finance, Anti-fraud, Anti-money Laundering (AML), Deep Neural Networks, GPU Computing, Sequence Models, Variational Autoencoders, Prototyping, Artificial Neural Networks (ANN), Classification, Regression, Predictive Analytics, Supervised Learning, Classification Algorithms, Decision Trees, Decision Tree Classification, Decision Tree Regression, Regression Modeling, Random Forests, Gradient Boosted Trees, Software Development, Solution Architecture, Clustering, Clustering Algorithms, High Code Quality, Solution Design, Technical Product Management, A/B Testing, Agile Sprints, Data Science Product Manager, Competitive Programming, Object Detection, APIs, Analytics, Research, Computer Vision Algorithms, Convolutional Neural Networks (CNN), Neural Networks, Financial Data, Financial Modeling, Scripting, Automated Data Flows, Signal Processing, Digital Signal Processing, Image Processing, Models, OCR, Data Visualization, Big Data, Big Data Architecture, Video Processing, Health, Video Analysis, Back-end, Datasets, Simulations, Text Generation, Google BigQuery, Speech Recognition, Startups, Early-stage Startups, Public Speaking, Voice Analysis, AI Programming, Chatbots, Minimum Viable Product (MVP), Proof of Concept (POC), Education, Contract, Hugging Face, Reinforcement Learning, Financial Forecasting, Image Generation, Financial Engineering, Software Architecture, Technical Leadership, Full-stack, Advisory, Project Consultancy, Data Analytics, CTO, Leadership, Text Recognition, Software as a Service (SaaS), GPT, Generative Pre-trained Transformers (GPT), OpenAI GPT-3 API, Hedge Funds, Recurrent Neural Networks (RNNs), Explainable Artificial Intelligence (XAI), OpenAI GPT-4 API, Text to Speech (TTS), Speech to Text, Frameworks, Generative Pre-trained Transformer 3 (GPT-3), Workshop Facilitation, Data-driven Decision-making, Decision Modeling, Expert Systems, DSP, Voice Recognition, WebAssembly

Collaboration That Works

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

Discuss your requirements and refine your scope in a call with a Toptal domain expert.

Choose your talent

Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.

Start your risk-free talent trial

Work with your chosen talent on a trial basis for up to two weeks. Pay only if you decide to hire them.

Top talent is in high demand.

Start hiring