Machine Learning Engineer with 5+ years of experience in Deep Learning and Machine Learning, including Natural Language Processing, Time Series and Computer Vision areas. I build custom solutions that fit most the particular need. I love to work in NLP and CV fields.
Languages: Python, C++
Frameworks: Tensorflow, Keras, scikit-learn, SciPy
Data manipulation: Pandas, Numpy
Image processing: OpenCV, scikit-image
ML Engineer, Tbilisi, DataArt, 2021 - present
The project aimed to develop an optimal energy consumption system for residential clients. The system was comprised of several micro-services that communicated via gRPC and Kafka, with data stored in TimescaleDB. Implementing, integrating, deploying, and improving some of these services. Conducted an analysis to determine the potential savings and the accuracy of different time-series data predictions for these services.
Developed a two-stage service for electric vehicle smart chargers. The first service predicted the plugged-in time and consumption, along with the expected errors for these values. The second service made a decision based on factors such as the available time slot, confidence, and other factors, ultimately generating an optimal schedule or forcing immediate charging.
- Data cleansing, grouping, and exploratory analysis
- Analysis of various cost optimization strategies
- Creation of simulation environments for the development of optimization
algorithms and hypothesis testing
- Development and improvement of several services, deployment, testing, and
- Training of various ML models ranging from simple Random Forests to combined Neural Networks with custom techniques.
Tools: Pandas, sklearn, PyTorch, CVXPY, DVC, MLFlow, MyPy, poetry, Streamlit, Docker, k8s, Protobuf, Kafka, RxPY, Prometheus, Grafana, Bamboo.
ML Engineer, First Line Software, 2020 - 2021
- Formulation of business problems in terms of ML (ML task, metrics, requirement and bounds).
- Creation of PoC (Proof of Concept), projects estimation, help in proposal forming.
- Research of modern approaches to machine learning and deep learning problems solving, development of these approaches.
- Writing applications for scientific grants, writing reports for the completed grants.
- Construction and maintaining of ML pipelines.
R&D Deep Learning Engineer, NNFormat, 2019 - 2020
- Data analysis, building models and automatic DL systems
- Research and verification of optimal architectures of decision functions (mainly NN)
- Development of solutions for daily forecasting of electricity consumption by enterprises and various electricity prices, making of energy consumption forecasts taking into account the price forecast. Costs and penalties were reduced on average by 2 times, for a long period of time (several months).
- Teaching data structures and algorithms, data science in Python.
Tools: Python, Bash, PyTorch, XGBoost, H2O, HpBandSter, Pandas, NumPy, Scikit-learn, Matplotlib, requests, PyCharm, Linux, JupyterLab
Software Engineer, GraviLink LLC, 2018 - 2019
- Supported and developed the functionality of a mining program.
- Worked on accelerating the mining process for various cryptocurrencies on diverse hardware architectures and video cards.
Tools: C/C++, OpenCL, boost, CUDA, Windows, Microsoft Visual Studio, Linux, MacOS, Docker
Full-stack Developer Intern, SPSU, 2017 - 2018
Developed a site for students and teachers, added pagination and various features.