Georgia Institute of Technology

Franklin University

Northern Virginia Community College

Lead Software Engineer

Clarence Health

Senior Engineer

• Enabled Customers to Search and Chat with their Data: Developed RAG (Retriever Augmented Generation) pipelines and integrated with OpenAI interfaces, enabling customers to conduct semantic searches and chat interactions over various data sources.
• Fine-Tuning of GPT LLMs: Fine-tuned OpenAI's GPT Large Language Models to specific datasets, enhancing model relevance and accuracy.
• Building Fine-Tuning Pipelines: Established fine-tuning pipelines, encompassing data acquisition, preprocessing, synthesis, hyperparameter tuning, and integration with the fine-tuned models.
• Troubleshooting RAG Pipelines: Diagnosed and resolved issues within RAG pipelines, including text embedding, storage in vector databases and querying for document retrieval to enrich LLM prompts.
• Enabling Chat over JSON documents: Used OpenAI's Assistant API to enable customers to effectively query their JSON documents
• Optimized for Performance: Improved data processing speed by vectorizing code using Pandas.
• Cloud Provider Code Abstraction: Abstracted cloud provider-specific code to facilitate easy swapping
between implementations, enhancing system flexibility and interoperability.
• Azure Integration: Successfully implemented integrations with Azure File Storage and Azure Key Vaults, transitioning away from AWS S3 and AWS Secrets Manager.
• OAuth Integration Troubleshooting: Diagnosed and resolved OAuth integration issues, ensuring secure and seamless system access.
• Maintaining and Updating OpenAI Plugin Setup: Consistently maintained and updated OpenAI plugin configurations, ensuring functional integration with ChatGPT.
• Python, FastAPI, Pydantic, MyPy, Mongo, Redis, Pinecone, Ray, Azure, AWS, Git, GitLab, Postman, Haystack, LangChain

Mantium

• Contributed to the success of a cybersecurity AI product by improving its functionality, expanding its capabilities, and ultimately leading to its sale.
• The product used machine learning and statistics to analyze logs, namely
• Hidden Markov Models, R, MGCV’s GAM, to model and weight normal versus threat activity
• Assisted GoSecure in servicing a diverse client base across various industries and domains. Some examples:
• Wrangled and modeled time series data using R, tidyverse, xgboost, python, scikit-learn
• Contributed to development of backend microservices - Java, Spring, docker, k8s, Rest, Postman, PostgreSQL
• Converted SAS scripts to Big Query stored procedures
• Automated SQL analysis through Selenium, JSQLParser and SQLite
• Built and troubleshot data pipelines - Pulsar, Logstash, Elastic, Kibana

GoSecure

AI Engineer

• Empowered both customers and engineers by designing and building scalable web applications from the ground up, enabling efficient service of distribution centers.
• Ensured Vargo's long term success through ongoing communication and knowledge sharing with fellow engineers and management.
• Java, Python, Spring, REST,SQL, MySQL, RabbitMQ, Tomcat, Linux, Maven, Cucumber
• HTML, CSS, SASS, Bootstrap, Javascript, Handlebars.js, jQuery, Websockets, Gulp

Vargo Companies

• Helped Franklin build features for a Learning Management System (LMS)
• Improved critical performance by redesigning and re-implementing analytics processing of the LMS
• Facilitated smooth integration with customers' authentication systems by redesigning and re-implementing
authentication module of the LMS
• Assisted in the integration of the company's systems with third-party software vendors through the
organization of joint sessions and the development of proof-of-concepts
• Java, Spring, JUnit, Javascript, jQuery, REST, MySQL, Tomcat, Linux

Boyko Todorov