Giorgi B.

DX Compliance Solutions is a UK-based FinTech company specializing in financial crime detection and prevention.

• Developed and trained machine learning models to detect fraudulent transaction patterns using large-scale synthetic datasets, improving anomaly detection accuracy.

• Built graph-based data representations in Neo4j to uncover hidden connections between entities, supporting detection of complex financial crime typologies.

• Collaborated with backend developers and data engineers to optimize data ingestion pipelines, ensuring efficient model training and feature generation.

• Enhanced model explainability by implementing network visualizations and SHAP-based interpretability reports for compliance review.

• Tuned and validated ML pipelines to balance precision and recall, reducing false positives while maintaining regulatory-grade accuracy.

• Deployed and monitored ML components in production, ensuring data consistency and proactive detection of model drift.

TuringofTrust is an early-stage IT startup building AI-driven products leveraging Large Language Models to deliver intelligent, context-aware applications.

• Designed and implemented a Retrieval-Augmented Generation (RAG) architecture using LangChain and LlamaIndex, integrating vector databases for semantic search.

• Developed modular Pydantic object structures to streamline LLM agent interactions and ensure schema validation across data pipelines.

• Integrated OpenAI APIs with Azure cloud components, deploying MVPs as containerized Dash applications for beta testing.

• Optimized data embedding and retrieval workflows using Azure Cosmos DB for MongoDB and PostgreSQL backends.

• Deployed the full prototype to Azure App Service and implemented Docker-based CI/CD for scalable testing and iteration.

Lead development of real-time ad fraud detection and behavioral ML systems using large-scale user and traffic data. Build and deploy production ML models, backend services, and data pipelines using Python, GCP, BigQuery/Bigtable, PostgreSQL, and RabbitMQ. Work on fraud, bot, VPN/proxy, behavioral and conversion detection, as well as model monitoring and experimentation. Own solutions from data analysis and modeling through production deployment and optimization.

• Developed and implemented NLP (Natural Language Processing) models.
• Utilized Python for machine learning and data processing tasks.
• Worked with TensorFlow and PyTorch for model development and training.
• Implemented machine learning solutions on AWS using services like EC2 and S3.
• Managed databases with PostgreSQL.
• Built APIs using FastAPI for application integration.
• Handled asynchronous tasks with Celery.
• Integrated and leveraged the ChatGPT API for advanced NLP functionalities.
• Employed RabbitMQ to handle message queuing requirements.

• Implemented neural language processing techniques to optimize the hiring process and improve employee referrals by matching job requirements with suitable candidates from employee LinkedIn contacts.
• Trained and deployed deep learning models to enhance the accuracy and efficiency of data-driven decision-making in the employee referral program.
• Conducted data scraping and processing to curate relevant datasets for training deep learning models, contributing to the development of sophisticated algorithms and solutions for talent acquisition.
• Proficiently deployed Python services using Flask, FastAPI, Celery, RabbitMQ, and Cronjobs for seamless integration and process automation.

• Leveraged deep learning frameworks such as PyTorch and TensorFlow to engineer advanced algorithms that accurately identified and tracked customers, thereby enhancing the overall customer experience within the cloud-based POS and business management system.
• Adeptly managed databases using PostgreSQL and MS SQL to store and retrieve crucial data, ensuring data integrity and efficient access.
• Demonstrated a dynamic combination of computer vision engineering, Python development, and machine learning expertise at Franpos, resulting in a robust facial detection and tracking system within the cloud-based POS and business management system.
• Trained and deployed deep learning models to enhance accuracy and efficiency in data-driven decision-making for the employee referral program.
• Conducted data scraping and processing to curate relevant datasets for training deep learning models, contributing to the development of sophisticated talent acquisition algorithms and solutions.
• Proficiently deployed Python services using Flask, FastAPI, Celery, RabbitMQ, and Cronjobs for seamless integration and process automation.

• Developed and deployed an Automatic License Plate Recognition (ALPR) system using PyTorch, TensorFlow, and Keras;

• Implemented a Body Poses Estimator for Golf players utilizing deep learning techniques;

• Created a Nudity Detection system for online social media platforms and executed data mining and scraping tasks with Python and popular libraries such as Scrapy, Beautiful Soup, and Selenium;

• Proficient in cloud platforms like AWS and GCP for efficient model deployment and management;