Tomasz Tkaczyk

About Tomasz

Experience

1. F

   Freelance

   • AI & ML Consultant

     Jan 2023 - Present · 2 yrs 9 mos·Remote·Current

     Consulting in the domain of GenAI and LLMOps, delivering custom AI app development services with a focus on advanced language models. I provide strategic consulting and hands-on development, transforming ideas into feasible solutions. My expertise includes: AI Strategy & LLMOps: Advising on AI integration and operational strategies, ensuring scalable and efficient implementation of language models in business processes. Custom AI Applications: Developing bespoke AI solutions, from GPT-based data tagging API to advanced RAG systems leveraging multiple fine-tuned LLM agents. Proof of Concept (PoC) Development: Spearhead the development of PoCs to demonstrate the practicality and potential impact of AI solutions in real-world scenarios. My role as a consultant goes beyond technical development and advisory; I am a strategic partner to my clients, helping them navigate the complex landscape of AI technology to build AI-powered tools that solve real problems.

2. D

   DoiT International · Full-time

   • Software Engineer - Data & AI

     Aug 2021 - Present · 4 yrs 2 mos·EMEA·Current

     - Delivered and maintained critical ETL pipelines processing ~10TB of AWS billing data per month (SQL, Apache Beam, Numpy) - Proposed the idea and led the implementation of customized internal dashboards/admin panels that automated countless hours of manual work (1/2 FTE), as they are used daily by a team of analysts (Plotly, Dash, Streamlit, FastAPI) - Designed, implemented, and monitored all stages of MLOps for the cloud usage forecasting pipeline. Used an ensemble of ARIMA and DNN to make daily forecasts for over 2,000 customers which improved the profitability of the entire core by at least 10%. (Kubeflow, Docker, Sklearn,Pytorch) - Led the team to win the company-wide Hackathon with innovative optimization idea that will save our customer ~ $800k in a year - Used Streamlit to build a pipeline monitoring dashboard which became the default ETL debugging/ troubleshooting tool within my team DoiT International, a strategic partner of GCP and AWS, tackles complex problems of scale for our customers, using our expertise in resolving problems, coding, algorithms, complexity analysis, and system design.

3. A

   Axel Springer SE · Full-time

   • Software & Data Engineer

     Aug 2020 - Aug 2021 · 1 yr 1 mo·Berlin, Germany · Hybrid

     - development of a Topic Detection and Tracking method combining NLP and deep learning with unsupervised document clustering - implemented a dozen Slack bots in JavaScript and AWS Lambda delivering data-driven insights to editorial teams While working full-time as a grad student I've finished and defended my master's thesis with the highest grade. Master's Thesis topic: "Design and evaluation of document embedding-based topic detection and tracking system for news articles"

4. I

   IBM · Full-time

   • Data Science Researcher

     Mar 2019 - Apr 2020 · 1 yr 2 mos·San Francisco Bay Area

     My main project was called: "Machine Learning-based Ransomware detection for Cyber Resiliency" • Successfully delivered anomaly detection ML models in Python for a cybersecurity application detecting 94% of anomalies. • Contributed to the design and implementation of an ML product for IBM's Recovery Orchestration. • Researched and implemented data generation pipeline using malware Sandbox (Cuckoo) to collect the data from over 200+ ransomware strains including CryptoLocker, Jigsaw, and WannaCry. • Lead the design of the PoC based on the evaluation of state-of-the-art research in the field of Cyber Security, Cyber Resiliency, Anomaly Detection, and AI. Technologies and stack: Sklearn, Pytorch, Redis / ActiveMQ, Docker, MariaDB, MinIO, Pandas, Dask, Kubernetes IBM Cloud + Watson, Cuckoo, VMWare

5. D

   Digitas Pixelpark · Work Study

   • AI/Data Science Working Student

     Aug 2016 - Feb 2019 · 2 yrs 7 mos·Berlin und Umgebung, Deutschland

     Technical: - successfully delivered a CMS application developed using Node.js, Angular & MongoDB to serve customized analytics dashboards to the clients - implemented over 20+ ETL pipelines in Apache Airflow & PySpark - manged AWS architecture and carried out DevOps jobs with CloudFormation, IAM, and ECS - developed three Voice UI's and Amazon Alexa Skills for internal use Analytics: - taking responsibility for preparing custom Tableau Dashboards for brands like i.e Mercedes-Benz, McDonald or BMW - provided functional support for data analysts in MongoDB/SQL queries, Google Analytics, GTM and A/B Testing with Optimizely - build over 10+ Tableau analytics dashboards provided to clients Stack: Tableau Desktop, Tableau Server, Apache Airflow, Pandas, Flask, Amazon S3 + Redshift, SageMaker, AWS Lambda Team Presentation

Education

1. 
   Freie Universität BerlinMaster of Science - MS, Computer Science & Media2018 - 2021

   Grade: GPA 3.7 / Endnote 1.3 Joint-Degree between Humbold, Technische and Freie Universität Berlin. Coursework: Machine Learning, Data Visualisation, Data Integration, Big Data Analytics, Data-driven Business, Enterprise Computing Master Thesis

2. 
   Technische Universität BerlinBachelor's degree, Computer Science and Media2014 - 2017

   Activities and societies: - Mentored younger Medieninformaitk colleagues as a part of the faculty-wide program Joint-Degree between Technische and Freie Universität Berlin. Coursework: Algorithms and Data Structures, Web technologies, Linear Algebra, Advanced Multivariate Calculus, Statistics, Media Science Bachelor Thesis: "Design, Development, and Evaluation of a Voice User Interface for Queries on semistructured Data"

Skills