Michał Martyniak

Responsibilities:

• leading the process of datasets versioning tool standarization (RFC),
• feature store selection process for client's AI department (RFC),
• occasionally interviewing candidates for SE in ML roles,
• developing a SageMaker pipeline template and more

• Python
• AWS: SageMaker Pipelines, CodeBuild, CodePipeline, IAM, EC2, S3, EMR, Secrets Manager
• IaaC: Terraform
• Feature Stores: Hopsworks and Tecton
• DVC for data versioning
• Neptune.ai for experiment tracking
• Github, Jira, Miro

Building a custom engine for parsing natural language expressions with TDD approach. Besides that I had a chance to do an internal presentation about effective terminal workflows and tools. Stack:

• Python
• Java
• ANTLR

We have transferred the data of real-world car traffic to the simulation. The released scenarios cover maneuvers like overtaking on freeways or exiting roundabouts. This research and the extensive set of experiments is a result of collaboration with a big car manufacturer. In short:

• Main author of the bird-eye's visualization library on PyPI (Github repo)
• Recreating real-world maps for CARLA (UE4) using RoadRunner
• Co-author of an article accepted to the Autonomous Driving Workshop at NeurIPS 2020 (arxiv version)

• Python (Tensorflow, NumPy)
• CARLA + RoadRunner
• Neptune.ai for experiment tracking

NLU-based chatbots and voicebots development, multi-label text classification with fastai Attempts to introduce reproducible ML pipelines and experiment tracking methodology Reading ASR-related papers Stack:

• Python (FastAPI, SQLAlchemy, fastai)
• Docker
• DVC and Guild AI for experiment tracking

Highlights:

• Co-author of a paper published in JMLR (Journal of Machine Learning Research)
• Designing a modular architecture and REST API
• Incorporating parallel-ssh library for multi-node communication
• Implementing GPU monitoring module using custom-written nvidia-smi parser
• Adding GPU reservation mechanism with resource protection
• Introducing automated tests, database migration etc.

• Python 3.5+ (pytest, Flask, SQLAlchemy, Alembic, parallel-ssh, flake8)
• Swagger UI/OpenAPI (interactive documentation of endpoints)
• Git with GitFlow strategy

It consisted of 4 important parts:

• Dashboard for administrators - Web app for managing users and their access, additionaly WYSIWYG editors, email sending, ACL logs
• Web interface for clients - displays responsive, platform-specific content (iOS, Android, PC) prepared by an admin in WYSIWYG editor
• Linux daemon - my own iptables manager written in Ruby, which controls the life of every rule and talks to the database periodically
• Database project

• How to containerize Rails applications with Docker (docker-compose), how neat it is in deployment and maintainance
• Managing my time
• Customizing gem's source code
• Ruby Metaprogramming
• That automating my work really matters
• What real-world clients might expect from the end product

I've learned a little bit about:

• SMTP
• Bash and Perl scripting
• My fear from PHP syntax

Additionaly we were working with refactoring hardware's firmware written in C. I've learned a lot there, including:

• Git
• MVC architecture
• Rake, Active Record, templating html and more