Machine Learning Engineer / Data Scientist

Experience

Data Scientist, RBC Group, 10/2020 - 12/2020
Machine Learning Engineer, Adoriasoft, 02/2021 - 10/2021
Machine Learning Engineer, FreySoft, 11/2021 - 12/2021
NLP Researcher, Team Lead, Sandsiv+, 01/2022 - 04/2023
Consulting, freelance, 07/2023 - ...

I now have almost 3 years of experience in building ML/DL pipelines with 2.5 years of commercial experience in the field including team-leading experience.

Currently providing occasional consulting services, e.g. on the RAG project. My latest job was aimed at developing various textual NLP solutions (incl. ABSA) for the VoC platform allowing our clients to better understand their customers' feedback.

My previous position involved developing personalized marketing solutions using various NLP methods. Before that, I was working on DL-based steganography (undetectable watermarking) with different content modalities, primarily employing Computer Vision techniques.

I have experience applying LLMs and GANs to solving custom problems. In my work, I leveraged the prompting of existing models and the development of custom ones with specific training procedures like off-graph usage of non-differentiable computation blocks or Triplet loss for Siamese nets.

I'm familiar with popular frequency transforms (DCT, DFT, DWT) and know Reinforcement learning basics (DQ learning).

I have publications in professional scientific journals, in particular on the topic of ML/DS, e.g. I've proposed a way of dealing with the Peter principle in hierarchical organizations via using machine learning.

I have a general understanding of DeFi and have implemented interpretable ML-based credit scoring systems, including via SHAP values analysis.

Skills

Machine Learning Python Data Science Deep Learning numpy Keras Pandas English NLP PyTorch PostgreSQL AWS Docker Git Team Lead LLM ChatGPT Rasa LlamaIndex

Highlights

1. Developed a method for the embedding of the non-differentiable computation blocks in the training pass of the NNs without breaking the gradient flow. The method has boosted the project and was employed by other engineers.

2. Developed a novel approach to the ABSA which involved combining several solutions that were non-conventional/not intended for this task but allowed for wider and more flexible analysis. That helped our clients better understand their users' feedback by gaining a more nuanced look due to the power of DL.

3. Successfully led a junior team member towards professional growth and accommodating the specifics of the NLP R&D process.

Looking for

I'm looking for a job that would allow me to broaden my ML/DL skills and realize my interests.
What I don't want is ML for the sake of ML and applying AI simply as a hyped instrument.

Preferred language

Українська, English