Creating knowledge base using AI toolchain (embeddings, gpt) for internal engineering organization (OpenAI, LangChain, LlamaIndex)
Building API platform based on GraphQL, Apollo Federation, and Service Mesh / Envoy (TypeScript, Gradle, Jenkins, GitLab CI/CD, AWS, Terraform, Spark, Java, Kotlin, Rust)
Leading team of contractors to create a new backend service, fully integrated with existing systems to provide new capabilities on employer portal (Java, Kotlin, Gradle)
Implementing new and adjusting existing client-facing products for enterprise clients on employer portal for better content management over job postings, ads campaigns, and candidates messaging (Java SE/EE, Kotlin, JavaScript, React, Python)
Key accomplishments:
Built a chatbot (standalone and Slack integrations) to address internal customer requests
Indexed thousands of Confluence pages, Google documents, emails, and Slack threads using OpenAI embeddings
Using LangChain preparing prompts for GPT-4 to generate answers with references to relevant resources
Currently improving performance with migrating to the LlamaIndex stack
Contributing to GraphQL Gateway internal platform
Supported migration and onboarding of 250 services to GraphQL stack and over 1000 API clients
Led continuous performance evaluation and improvements, currently able to support billions of daily requests and 40,000 peak RPS with over 600 globally distributed nodes
Co-piloted staged intricate migration to Apollo Federation version 2
Personally led migration to a new API Gateway product vendor (from Apigee Edge to Cloudflare)
Conducted feature research for all available solutions on the market
Prototyped solutions and performed performance and load testing
Participated in contract negotiations
Built end-to-end logging infrastructure
Resulted in $1M (80%) annual cost savings, improved reliability and performance metrics, and richer telemetry
Created important components designs
Data durability contracts (similar to Microsoft) fitting company’s environment
Client fine-grained authorization system leveraging GraphQL DSL, OAuth, and Open Policy Agent’s authorization DSL
Contributing to various engineering organization initiatives
Supporting and delivering new features to web crawler ecosystem for aggregating job postings from all over the Internet (Java SE/EE, JavaScript, Python)
Improving automatic crawler rules repair tooling using heuristics and machine learning
Key accomplishments:
Personally delivered crowdsourcing platform for human intelligence tasks
Collecting requirements, communicating with stakeholders
Holding system design review, prototyping, and writing documentation
Releasing it to the clients, onboarding new engineers
Resulted in reduction of thousands of hours saved, hundreds of thousands tasks completed, assisting in bringing over $10M in new customers LTV
Supporting migration of crawling infrastructure to industry standard web browsers using Selenium and Selenium Grid (clustering solution)
Preparing and consolidate data from various storages and services to enable data scientists to experiment and come up with ideas of fully automating aggregation pipeline
Personal project of leveraging publicwww.com dataset for building websites classifier to generate leads for discovery system in the crawler, resulted in 750k positively identified websites
Participating as an engineering lead in Indeed University 2017, internal startup incubator program to on-board new engineers and generate/test new ideas for business, acted as an interim manager for a cohort of 5 people
Mentoring interns and junior team members, hosting internal tech talks and workshops about build tools, testability, web app technologies, conducting technical interviews for SWE/TDM candidates
Developing and supporting web analytics tool Yandex.Metrica (Java SE/EE)
Key accomplishments:
Whole support of API component, introducing new reports functionality
Working closely with data backend counterpart team, contributing small fixes to now open source project ClickHouse
Working closely with QA automation engineers to create automated regression testing of all documented API endpoints based on automatically generated JSON schemas
Setting up thorough stress / load testing of beta versions before deployment, leveraging Yandex.Tank internal cluster
Developing and supporting game mechanics of MMORPG “Allods Online” (Java SE, C++)
Key accomplishments:
Developed several new game mechanics (abilities tied to distance between coupled in-game objects)
Fixed client reported visual bugs which were in game for years (C++, Lua)
Identified a bug with texture mapping for multiple matrix operations, causing numeric precision loss, fixed with simplifying to a single matrix multiplication
Built unit testing harness for brand new project based on Unity Engine