# Bogdan Kapusta

> Senior Software Engineer, Search | Lead Search Engineer | Vector Search & ML Specialist | RAG | AWS OpenSearch | Elasticsearch | SOLR | LLM | Cost Optimization & DevOps

Location: Santa Clara, California, United States
Profile: https://flows.cv/bogdankapusta

●	Architected and led large-scale Vector and Keyword Search systems on AWS OpenSearch and Solr for 300+ million records.
●	Performed exploration of best available models for Vector Search.
●	Leveraged LLMs (OpenAI, AWS Bedrock) for search relevance analysis.
●	Developed GPU-accelerated vectorization pipelines using Databricks to process 100M+ documents.
●	Created backend applications using Java/Python/React JS/Docker/ on AWS ECS.
●	Contributed to Terraform templates in the company (AWS OpenSearch, ECS, EC2).
●	Optimized cloud infrastructure, reducing costs by $240K/year through sharding and engine selection.
●	Proposed external applications integration (Discord, Telegram, WhatsApp bots) with internal APIs for enhanced user engagement.
●	Contributed to company-wide architectural forums, influencing technical direction and best practices.
●	Influenced a company-wide shift in understanding of user search behavior.
●	Fluent in English, Ukrainian, and Russian; strong cross-team mentor and collaborator.
●	US Citizen.

## Work Experience
### Staff Software Engineer @ Uniphore
Jan 2025 – Present | Palo Alto, California, United States
Agentic AI

### Senior Software Developer @ Chegg Inc.
Jan 2019 – Jan 2025
●	Architected and led implementation of a high-scale vector search over 100M+ records, achieving 2x latency reduction and 2.35% lift in core relevance metrics; served as foundational infrastructure for enabling Retrieval-Augmented Generation (RAG) systems.
●	Participated in company-wide architecture decision records meetings.
●	Conducted offline evaluation of search relevancy using OpenAI’s batch API and AWS Bedrock LLMs 
●	Built GPU-accelerated vectorization pipelines in Databricks, efficiently processing 100M+ records for large-scale ML workloads.
●	Performed hardware cost evaluation and optimization for OpenSearch reducing infrastructure expenses by $240,000 annually.
●	Benchmarked hybrid search, revealing 38% CPU usage at 100 RPM with only 4% NDCG gain; recommended deprioritization based on cost-benefit analysis.
●	Created scalable applications using Python, Spring Boot, Docker, deployed on AWS ECS for seamless scalability and performance.
●	Analyzed A/B testing results to uncover user behavior patterns and inform relevance tuning. 
●	Influenced a company-wide shift in understanding of user search behavior by formalizing match types (exact, beneficial, bad). Which caused enabling a classifier that achieved ~2x F-score improvement over the baseline and reduced LLM usage, leading to operational cost savings.
●	Proposed external applications integration (Discord, Telegram, WhatsApp bots) with internal APIs for enhanced user engagement.
●	Authored and maintained Terraform templates supporting infrastructure as code for OpenSearch, EC2, and ECS.
●	Coordinated with AWS and Hugging Face SMEs to leverage advanced ML tools and ensure industry best practices.
●	Mentored engineers across teams on RAG-related projects, supporting cross-functional collaboration.
●	Mentored interns on the Search Engineering Team, fostering technical growth and contribution.

### Senior Software Developer in Test @ Chegg Inc.
Jan 2018 – Jan 2019
●	Automated 500+ test cases in Robot Framework + Python environment
●	Performed offline comparison of search engines using Pandas, Numpy, scikit learn, fasttext libraries in Jupyter Notebook (3000+ lines of code)
●	Designed REST API automation for SOLR and microservices’ endpoints
●	Deployed Data Science microservices in AWS ECS
●	Performed offline testing analysis for search engine precision optimisation using DS model.
●	Extracted data from Redshift database and parsed it with pandas library
●	Enhanced testing framework with APIs for MongoDB interactions
●	Created application that returns prediction of similar topics based on input string using (React JS) and BE (using Python’s Falcon and gensim). BE was hosted using Docker container
●	Integrated support of Chrome Extension automation for testing framework using image recognition library
●	Added automation in jenkins pipelines for production deployment verification
●	Created load tests using Gatling for simulating production traffic in test environment
●	Established dockerfiles for building images with enhanced functionality
●	Performed deployment to production environments with following assertion of post deployment stability using Logz.io and New Relic applications
●	Set up alerts in New Relic for production environment
●	Participated in team that took second place in Hackathon 
●	Took part in multi continental meetings (Israel, India, Ukraine) with Project Owners, Developers, UX designers, providing suggestions for stories implementation 
●	Performed PR reviews of QA engineers and developers  in GIT
●	Created documentation to formalize release process in subteam
●	Created documentation for production environment setup
●	Interviewed candidates for QA Engineer position

### White Box QA Engineer @ PlayStation
Jan 2017 – Jan 2018 | San Francisco, California
●	Tested PS4 framework applications and web-based applications
●	Automated test cases using Python and Node JS.
●	Used Docker, MySQL and Shell scripts to simulate production environment based on Ember.js 
●	Participated in meetings with Project Owners, Developers providing suggestions for stories implementation
●	Performed PR reviews of QA engineers in GIT 
●	Used REST API for data population
●	Performed on-boarding trainings for software engineers. Created video walkthroughs for most prevailing features
●	Filed bug reports. Negotiated bugs with beta testers 
●	Documented automation software setup

### QA Automation Engineer @ Bitglass, Inc.
Jan 2015 – Jan 2017 | Campbell
•	Tested web-based cloud security application that provided security Office 365, Google Apps, SalesForce, Exchange, 
•	REST API, Soap API testing with Python
•	Experience in Performance test automation;
•	Programming in Google script 
•	Uploaded automated test cases to Jenkins using GIT repository
•	Automated test runs in Robot Framework using Data-driven and Behavior-driven styles
•	Managed Jenkins
•	Mobile Test Automation with Appium 
•	Tested on mobile devices including Samsung Galaxy tab 4, iPad mini, iPod, Windows RT
•	Created test plan, test cases, test runs for stories assigned to me at Bugzilla bug tracking system
•	Added command in Task Manager on Windows to run Ccleaner automatically when load on machine that was using Jenkins was minimal
•	Used Secured Shell to access log data
•	Used HMA! pro VPN application to simulate signing in to the developed product from different locations
•	Performed testing from remote virtual machines 
•	Used AD FS 2.0 to create users and verified population of user detail settings via SAML (attributes
•	Reported bugs and included log data in them
•	Negotiated with PM enhancements for various types of stories
•	Provided QA intern training

### QA Automation Engineer @ UkrVideo.com
Jan 2013 – Jan 2015 | Ukraine, Kyiv
•	Created and executed test cases for the flash player on a web-based application Ukrainian Movies Portal (video content collection with subtitles such as movies, TV series, cartoons, sporting events, tutorials, etc.) providing video-on-demand streaming services 
•	Automated 300+ test cases using Python, Selenium Webdriver
•	Tested Performance of website using Jenkins and Locust
•	Performed exploratory testing of the application to familiarize with the applications basic features
•	Tested many functional modules and product features in Agile environment
•	Participated in cross-browser (MSIE, Firefox, Chrome, Safari) and cross-platform (Windows 8/7/XP, MAC) testing 
•	Conducted regression and release acceptance testing for all the major and transitional releases
•	Participated in usability testing of the application and workflow involving multiple functional modules and features
•	Executed regression testing as defined by the project testing activities for a new QA build/release
•	Analyzed software failures, reported issues to Bugzilla
•	Administrated Bugzilla bug tracking system
•	Took responsibilities as team lead in test case development
•	Participated in regular project team meetings to discuss testing process, business requirements, and assignments
•	Used HMA! pro VPN


## Education
### Bachelor of Engineering - BE in Automotive Engineering Technology/Technician
Donetsk National Technical University


## Contact & Social
- LinkedIn: https://linkedin.com/in/bogdan-kapusta

---
Source: https://flows.cv/bogdankapusta
JSON Resume: https://flows.cv/bogdankapusta/resume.json
Last updated: 2026-04-12