Hi, I'm Szymon

I'm a machine learning engineer, with a focus on GenAI & Leadership

Get in Touch

If you'd like to get in touch feel free to contact me below!

I am a Machine Learning Engineer with over three years of experience in managing teams and leading projects and over 7 years of experience in developing in Python. Since founding my company, I have completed over 10 international projects and two patents were filed based on my work. My industry experience combined with a strong background in neuroscience makes me a perfect addition to any company working on cutting edge Machine Learning (ML) technology.

Beyond my technical skills, I also have 3 years of leadership experience at Director/C-suite level. I have also applied for and led government funded research projects, enabling me to confidently manage large-scale projects.

Work

2020 - current

Founder & CTO

OctoShrew Ltd.

OctoShrew is an ML Consultancy in which I am managing a team of Machine Learning Engineers to deliver solutions in the startup space. This challenges me frequently to complete ambitious projects under tight budget and time constraints. I am also responsible for making any hiring decisions and meeting the company’s financial targets.

Python
Pytorch
Tensorflow
LLMs
Generative AI
AWS
2020 - current

Founder & CTO

OctoShrew Ltd.

OctoShrew is an ML Consultancy in which I am managing a team of Machine Learning Engineers to deliver solutions in the startup space. This challenges me frequently to complete ambitious projects under tight budget and time constraints. I am also responsible for making any hiring decisions and meeting the company’s financial targets.

Python
Pytorch
Tensorflow
LLMs
Generative AI
AWS
2021 - 2023

Co-Founder

Coeur.ai

Coeur.ai is a SaaS company delivering Machine Learning based SEO services. I supervised and organized the development of the product from ideation to launch. The main product is a program that enables near autonomous writing of blog posts with minimal user input, reducing writing times for large-scale writing from ca. 2-3h to under 5 minutes per blog.

Python
Pytorch
Tensorflow
LLMs
Generative AI
AWS
2021 - 2023

Co-Founder

Coeur.ai

Coeur.ai is a SaaS company delivering Machine Learning based SEO services. I supervised and organized the development of the product from ideation to launch. The main product is a program that enables near autonomous writing of blog posts with minimal user input, reducing writing times for large-scale writing from ca. 2-3h to under 5 minutes per blog.

Python
Pytorch
Tensorflow
LLMs
Generative AI
AWS
2020 - 2022

Lecturer

Quickstart.com

Designed and taught two courses on the topic of Artificial Intelligence and Data Science with ca. 100 hours of lectures for quickstart.com. The course is taught at different renowned American Universities like UC Santa Barbara, University of Texas & Colorado State University.

N/A
2020 - 2022

Lecturer

Quickstart.com

Designed and taught two courses on the topic of Artificial Intelligence and Data Science with ca. 100 hours of lectures for quickstart.com. The course is taught at different renowned American Universities like UC Santa Barbara, University of Texas & Colorado State University.

N/A
2017 - 2020

Freelance Machine Learning Researcher

Self-Employed

During my studies, I worked as a freelance Machine Learning Engineer. I focused on smaller projects within the NLP and Computer Vision space. I learned professional communication with clients, writing and specification of requirements and laid the foundation for the skillset and professional network required for founding OctoShrew.

Python
Pytorch
Tensorflow
Scikit-Learn
Numpy/Scipy
Pandas
AWS
2017 - 2020

Freelance Machine Learning Researcher

Self-Employed

During my studies, I worked as a freelance Machine Learning Engineer. I focused on smaller projects within the NLP and Computer Vision space. I learned professional communication with clients, writing and specification of requirements and laid the foundation for the skillset and professional network required for founding OctoShrew.

Python
Pytorch
Tensorflow
Scikit-Learn
Numpy/Scipy
Pandas
AWS

Patents

2022

Image based Product Search

Invented a system allowing users to search through a database based on product images & lists of products. This required innovative combinations of multimodal models and OCR to create the necessary embeddings for search.

Python
Multimodal LLMs
OCR
2022

Image based Product Search

Invented a system allowing users to search through a database based on product images & lists of products. This required innovative combinations of multimodal models and OCR to create the necessary embeddings for search.

Python
Multimodal LLMs
OCR
2021

Statistical Modelling Method for Blackjack House Advantage

Invented a method that allows for real-time tracking of a casino’s house advantage. It calculates a player’s value to the casino as well as catching “advantage players” for zillions.ai.

Python
Statistical Modelling & Simulation
2021

Statistical Modelling Method for Blackjack House Advantage

Invented a method that allows for real-time tracking of a casino’s house advantage. It calculates a player’s value to the casino as well as catching “advantage players” for zillions.ai.

Python
Statistical Modelling & Simulation

Projects

2022

Automated Analysis of Pharmaceutical Records

CamHCR

CamHCR is a company focusing on consulting multinational Pharma companies such as J&J, Pfizer etc. Octoshrew was called in as external consultants to help with improving business operations. We identified that a company-wide retrieval and analysis system for data from past reports could improve productivity by up to 20%. Due to the highly specific nature of texts, an embedding model was fine-tuned and combined with GPT-3 to be able to provide analysis of the records to the consultants on-demand.

Streamlit
LLMs
NLP
AWS
Docker
Python
FastAPI
2022

Automated Analysis of Pharmaceutical Records

CamHCR

CamHCR is a company focusing on consulting multinational Pharma companies such as J&J, Pfizer etc. Octoshrew was called in as external consultants to help with improving business operations. We identified that a company-wide retrieval and analysis system for data from past reports could improve productivity by up to 20%. Due to the highly specific nature of texts, an embedding model was fine-tuned and combined with GPT-3 to be able to provide analysis of the records to the consultants on-demand.

Streamlit
LLMs
NLP
AWS
Docker
Python
FastAPI
2022

Autonomous Drone Control Systems

Stealth Startup

The client was a company focusing on software solutions for autonomous drone flight planning. OctoShrew took over the development of a system which allows autonomous control of over 100 drones simultaneously over large areas of open space requiring highly optimized algorithms as well as rigorous GPU optimization. Extensive experience in working with LiDAR & Vision data to safely navigate drones through a dynamic environment was gathered.

LiDAR
FastAPI
NumPy
AWS
Docker
Python
2022

Autonomous Drone Control Systems

Stealth Startup

The client was a company focusing on software solutions for autonomous drone flight planning. OctoShrew took over the development of a system which allows autonomous control of over 100 drones simultaneously over large areas of open space requiring highly optimized algorithms as well as rigorous GPU optimization. Extensive experience in working with LiDAR & Vision data to safely navigate drones through a dynamic environment was gathered.

LiDAR
FastAPI
NumPy
AWS
Docker
Python
2021

Big Data Analysis/Recommender Systems

Studyly.com is one of Austria’s largest e-learning providers. I oversaw the development of a student evaluation and exercise recommendation system which used a variety of methods to push the boundaries of the SOTA. The core element of the project was applying advanced methods such as Reinforcement Learning (recommendation) and LLMs (evaluation) such that the results still remained explainable as required by Austrian law.

CI/CD
Reinforcement Learning
LLMs
Real-time Data Processing
2021

Big Data Analysis/Recommender Systems

Studyly.com is one of Austria’s largest e-learning providers. I oversaw the development of a student evaluation and exercise recommendation system which used a variety of methods to push the boundaries of the SOTA. The core element of the project was applying advanced methods such as Reinforcement Learning (recommendation) and LLMs (evaluation) such that the results still remained explainable as required by Austrian law.

CI/CD
Reinforcement Learning
LLMs
Real-time Data Processing