Maria — AI/Machine Learning Engineer

Work Experience

AI/Machine Learning Engineer/Team Lead (Demand intelligence platform)

Duration: 1 year

Summary:

• Creation of a fully automated service for various predictions of city load in UAE based on current events, weather, and historical data from multiple non-synchronized sources
• The platform integrates geographical data at multiple levels to provide accurate forecasting

Responsibilities:

• Manage the project and lead the team
• Develop parsing modules for all data sources and analyze the data
• Develop time series forecasting models based on historical data
• Design and create database for storing and continuous update of data
• Process geographical data covering different geographical levels (country, emirate, city, district)
• Build pipelines for asynchronous execution of different models

Technologies: Python, OpenCV, Tensorflow, Google Maps, AWS (SageMaker, Lambda, S3, EC2, Redshift), Spark, MLlib, GeoJSON, Shapely

AI/Machine Learning Engineer (Nosis AI Voice Bot)

Duration: 6 months

Summary: Development of a chatbot for polling a large number of people regarding their political preferences for a particular event in Australia, designed to handle various human behaviors and simulate real-human agent interactions.

Responsibilities:

• Manage the project and lead the team
• Design conversational flow for polling
• Implement various types of questions for different expected or unexpected human behavior
• Apply fallbacks for questions without expected replies
• Tweak agent settings to imitate a real-human agent
• Manage demos and large group tests, upgrading the flow based on results

Technologies: Python, Dialogflow ES, Twilio, nltk, spaCy

AI/Machine Learning Engineer/Team Lead (Genie Words Ads Generation)

Duration: 1 year

Summary: Development of an AI system generating contextual selling advertising headlines and descriptions for texts scraped from websites in Danish and English, incorporating advanced text generation and filtering techniques.

Responsibilities:

• Manage the project and lead the team
• Develop text generation system based on Markov Chains and large language models
• Improve service with filtering system using distance metrics like cosine similarity, Levenshtein distance, and fuzzy logic
• Include sentiment, polarity, and linguistic analysis
• Implement keyword extraction algorithms
• Create generation algorithms based on templates
• Add human-in-the-loop evaluation for continuous model improvement

Technologies: Python, OpenAI, Bert, Transformers, SpaCy, NLTK, Numpy, Scipy, AFINN, Markov Chains, Pandas, scikit-learn, PyTorch

AI/Machine Learning Engineer (Wie Label)

Duration: 6 months

Summary: Creation of a platform for scanning products, classifying them, detecting specific features, and generating marketing descriptions, including pre- and post-processing modules for asynchronous generation and filtering.

Responsibilities:

• Classify products in images and data based on customer-provided ontology
• Extract labels, fit, material composition of products
• Design database and data structure
• Generate product descriptions using OpenAI in German
• Create pre- and post-processing modules for asynchronous generation, processing, and filtering

Technologies: Python, OpenAI API, Pandas, NLTK, Spacy, Boto3, Botocore, Supabase, OpenCV

AI/Machine Learning Engineer (Porch)

Duration: 2 years

Summary: Contribution to an automated lead bidding system on auctions, focusing on investigating poor model performance, researching and processing large datasets, and improving modeling and data processing pipelines.

Responsibilities:

• Analyze and investigate poor model performance
• Create local databases using Docker
• Participate in database migration to Azure
• Form and filter datasets for training
• Develop and train new models
• Optimize codebase for training and evaluation
• Research and implement new features for modeling
• Re-create pipelines using Apache Spark

Technologies: Python, SQL, Azure, Databricks, Docker, Polars, Pandas, XGBoost, MLFlow, scikit-learn, Spark, PySpark, Numpy, Matplotlib, Seaborn

AI/Machine Learning Engineer (TradeSun)

Duration: 2 years

Summary: Automation of processing unstructured financial trading documents to extract necessary information for financial transactions, improving throughput, turnaround, and thoroughness of trade operations.

Responsibilities:

• Extract and identify key fields from documents to classify type and predict relevance
• Develop new features, algorithms, and models; update existing ones
• Improve evaluation process and invent business metrics
• Log and solve bugs in the service
• Create a demo for real-time information extraction with click-highlighted areas
• Continuously improve existing features and processes

Technologies: Python, Pytesseract, AWS (SageMaker, Lambda, Step Functions, DynamoDB, Textract, S3), Pandas, Numpy, BeautifulSoup, Boto3, Botocore

AI/Machine Learning Engineer/Team Lead (Mauka Digital Lead Generation)

Duration: 6 months

Summary: Development of a fully automated lead generation system to find qualified leads, place them in the sales funnel, and increase quarterly revenue through advanced text and speech analysis.

Responsibilities:

• Manage the project and lead the team
• Analyze and filter datasets
• Build custom text classifiers to analyze human speech for sales management
• Implement methods for context and text analysis including sentiment analysis, language detection, and part of speech analysis
• Design and build database structure

Technologies: Python, Spacy, NLTK, Pandas, scikit-learn, Numpy, Scipy, NRCLex, Tensorflow

AI/Machine Learning Engineer (Mojo)

Duration: 6 months

Summary: Development of a platform for cell distortion recognition in clinical trials to accelerate selection of high-quality sperm, improving accuracy and enabling more targeted fertility treatments.

Responsibilities:

• Create algorithm to detect spermatozoids in microscope video feed
• Classify items by mutations to select suitable samples for conception
• Detect position, movement, direction, and speed of spermatozoids
• Apply filters to improve tracking processes
• Log and solve bugs in the service

Technologies: Python, OpenCV, Numpy, Scipy, Tensorflow, Tensorboard, Scikit-image, YOLO

Education

• V. N. Karazin Kharkiv National University
  Master of Applied Maths
  
• V. N. Karazin Kharkiv National University
  Bachelor of Applied Maths