# Vitali Karpinchyk

> Staff Software Engineer, Meta AI Measurement & Quality

Location: San Francisco Bay Area, United States
Profile: https://flows.cv/vitalikarpinchyk

An experienced software engineer/architect building highly scalable web services for machine learning, high performance computing and big data processing.

## Work Experience
### Staff Software Engineer, Meta AI Measurement & Quality @ Meta
Jan 2020 – Present | Menlo Park, California, United States
- Meta AI Measurement & Quality: developing systems and algorithms to evaluate the performance of Meta AI models and to assess and improve the overall quality of MetaAI outputs 
- Monetization / Ads Ranking ENG Group: Ranking & Foundational AI: developed next generation ranking models and infrastructure to deliver ads across major Meta products (Facebook, Instagram, etc)

### Software Engineer, Machine Learning Infrastructure at AI/ML, Cloud Services and Maps @ Apple
Jan 2015 – Jan 2020 | Cupertino, CA
- AI/ML: building orchestration/scheduling layer of Turi ML compute platform used by the majority of Apple's machine learning teams.
- Maps (SPG): Built a full stack of tools and services managing vehicle routing for geo data collection
- iCloud: built pipelines and infrastructure for internal stateful services, serving hundreds of millions customers: Object store, Search, Queue
- iCloud: designed and built a distributed task execution service, similar to Amazon SWF. It is now used by multiple teams to run DevOps workflows for the services above.

### Software Development Engineer, Amazon Local Search / Amazon CloudSearch @ A9.com
Jan 2012 – Jan 2015 | Palo Alto, CA
Amazon Local Search Service (search for points of interest on in Amazon Maps and similar applications).
built/owned:
- Spelling correction service for Amazon Local/Geo Search
- Real time search results correction
- Localization for Germany & UK
- Search results quality evaluation

AWS CloudSearch (http://aws.amazon.com/cloudsearch)

I owned a kernel resource management component as well as several other key components.

Delivered Projects:
 - High availability solution for the service (Multi-AZ)
 - Improving scalability of the service
 - Increasing transparency of the system: tools and metrics.

### Sr. Software Engineer @ Complete Genomics
Jan 2008 – Jan 2012 | Mountain View, CA
I worked on building genome analysis pipelines, including development of computational cluster management software, data processing workflows and different genome analysis applications. The pipelines utilize a rich stack of different technologies and tools based on C++/BOOST; Enterprise Java; Python; GNU Perl/Awk/Sed etc. and control a multi-thousand-core HPC cluster.

### Senior algorithm developer @ SAP Labs
Jan 2006 – Jan 2008 | Palo Alto, CA
Office of CTO, project "New HAT":
Created frontend/ backend server communication layer and data access components for the New HAT architecture; was responsible for the integration and final tuning of the system created by our group which included controller development, development and tuning of GUI generators.
 
Advanced Technology group, project "ABAP Navigator":
Working on ABAP reverse engineering projects I designed and implemented the following components: 
Program Diff - advanced algorithm of source code matching based on hashed AST representation.
A bunch of multi-thread algorithms for source code pattern mining. The algorithms were designed to benefit from huge amount of memory and work on AZUL appliances.
Implementation of CloSpan algorithm used for Copy&Paste detection, standard and local LCS algorithms.
I designed system context management to support different system configurations of the ABAP Navigator system.
Tools:  Java, Eclipse, ANTLR, AZUL, C#, MS SharePoint Server, SAP A1S/AP Systems.

### member of technical stuff @ Magma Design Automation
Jan 2005 – Jan 2006 | Santa Clara, CA
Working on the cutting edge OPC project as a member of Mojave OPC team designed and implemented OPC geometry engine, including: layout manipulation algorithms based on graph layout representation; basic and scan-line geometry algorithms, boolean geometry operations; bucket-structure based 2D object search; regression test engines, used to create fast QA tests; C++/TCL-interface and exposure of the geometry engine functionality into TCL.
Tools: gcc, MS Visual C++, make, gdb, Tcl, Valgrind, CVS.


## Education
### Masters in Computer science
Belarusian State University


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

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