Hi, I'm Miki Pokryvailo

Miki Pokryvailo

I’m an energetic, easy-going software engineer that thrives on solving ill-defined problems proactively and autonomously. I think pragmatically and communicate transparently. I’m a home-renowned chef and an avid learner of psychology. I work at a stealth startup 🥷 , and blog about random stuff.

my photo
Taken in Chena Hot Springs, Alaska. February 2020, -25°F/-32°C.

Experience

Stealth startup, Remote

Founding Engineer

February 2022 — Present

Building cool stuff :)

MadHive, New York, NY

Backend Software Engineer

December 2020 — February 2022

  • Working on MadHive’s core API platform, “MadServer,” which is responsible for handling a large percentage of internal and user-facing queries. Using Go, GCP, BigQuery, Postgres, and Swagger.
  • Designed and independently implemented a service from scratch to continuously sync critical user data from MadHive’s central database to BigQuery. This service needs to ensure perfect data integrity, and the design decisions I made were critical in making this not just possible, but easy to implement and debug.
  • Independently came up with and implemented a solution to reduce payload size by 10x for one of MadHive’s slowest and most critical queries in its UI using protobuf field masks.

Confluent, Mountain View, CA

Backend Software Engineer | Platform Services Team (API and CLI groups)

August 2019 — December 2020

  • Worked on a core platform component, a highly automated API framework that’s leveraged widely across Confluent Cloud services. Use Go, Bash, OpenAPI, protobuf/gRPC, and Postgres.
  • Won 1st place at a Confluent hackathon by speeding up service deployments to our test environment by nearly 7x, and enabling developers to interactively debug their test environments. The previous debugging workflow consisted of adding print statements and redeploying services until the bug was found. It's likely that this workflow has made debugging cloud services an order of magnitude faster for many engineers at Confluent.

    A message my co-worker got from someone about my hack.
  • Developed a mechanism to measure integration test coverage for the Confluent CLI (command line interface), a tool used extensively by both Confluent and its customers. On my own accord, I generalized the solution, open-sourced it, and wrote an accompanying blog post.
  • Won the People’s Choice Award at Confluent’s first hackathon for developing an interactive shell feature that significantly improved the usability of the CLI. This was the only award at the hackathon granted democratically, demonstrating the demand and viability of our project. I was the engineering lead to ship this feature into production, and did so successfully.
  • Launched an initiative to improve the quality of the CLI in which I single-handedly completed a refactoring of several thousand lines of code. This significantly improved the flexibility and robustness of the codebase and fixed a number of critical bugs. I received an “engineering superstar” award for this after being at Confluent for only four months.
  • Implemented a new type of API key and related API key-based authentication that’s widely used in Confluent’s UI and APIs.
  • Integrated tools in CI pipelines to ensure backward compatibility was maintained across Confluent’s private and public APIs.
  • Advocated for and implemented API infrastructure utilities that are used across Confluent Cloud services to provide a consistent public API experience.
  • I signed my offer when Confluent had 400 people, joined 10 months later when it had 850, and did 3-4 interviews a week for two months, less than a month after joining.

Amazon, Seattle, WA

Software Engineering Intern | AWS Marketplace Search Team

May 2018 — August 2018

  • Owned the end-to-end system design and implementation of a proof of concept optimizing the scalability and maintenance of AWS Marketplace’s search platform using several graph databases.
  • Leveraged the Apache TinkerPop graph computing framework for complex graph queries.

TodayTix, New York City, NY

iOS Software Engineering Intern

June 2017 — August 2017

  • Implemented the home page for the company’s new “Discover” feature.
  • Completed a major part of an app-wide API migration.

Projects

End-to-End 3D Bézier Surface Reconstruction Using CNNs

July 2019 — August 2019

  • Wrote a paper on using Convolutional Neural Networks (CNNs) to reconstruct 3D Bézier Surfaces from 2D grayscale images.
  • Independently conducted background research, and designed and implemented neural nets using the Keras framework.

API Performance Grading System

February 2019 — May 2019

  • Built an automated API performance-grading system as an undergraduate TA of a cloud computing course (CSE 356).
  • The system orchestrated several machines to generate load, aggregate latency results, and grade the APIs accordingly.
  • Implemented the core logic of the system - the fuzzed load testing mechanism - from scratch. Implemented server-to-server communication using RabbitMQ, and designed some MongoDB schemas.

Cloud-Based High-Performance Social Networking API

March 2018 — May 2018

  • Implemented a social networking API in Go that authenticates, posts binary and text data, follows posts and users, and performs complex searches on tweets.
  • Designed a distributed microservice architecture and deployment automation with AWS EC2 and Ansible.
  • Used Nginx load balancing, a sharded MongoDB cluster, Elasticsearch, and a Memcached layer to achieve low latency query times under loads of hundreds of requests/s.

High-Speed Line Tracking Robot

January 2018 — February 2018

Independently assembled and programmed a robot using a Raspberry Pi, OpenCV, and Python, to autonomously follow a pattern. Implemented and optimized digital PID control to achieve high tracking accuracy and speed.

Skills

Languages

Proficient in Go, Java, and Python. Intermediate in Bash and JavaScript. Some experience with C, C++, and Lua. Currently, my favorite language is Go.

Infrastructure

Linux, Kubernetes, Postgres, Terraform, Vault, MongoDB, AWS (EC2, ELB, Neptune), Nginx/Apache, Elasticsearch, Memcached, Ansible, Protobuf/gRPC, OpenAPI, JSON:API, Node.js.

Other

Keras, neural networks (mostly CNNs), OpenCV, and PID control.

Education

Stony Brook University, Computer Science Honors Program

Double Major in Computer Science and Applied Mathematics • 2019

When not working

  • I'm a pretty decent home chef, a casual powerlifter, a hiker and runner, a guitar player and singer, and a lover of dark humor and stand-up comedy.
  • I love smoking meats (I make a mean rack of beef ribs) and cooking Mexican food. I can make tacos from scratch, including the tortillas.
  • I own a Martin X Series Custom X1-000 guitar (which I play badly), and love to sing anything from Sinatra to The Beatles, to Billie Eilish.
  • I'm working on cultivating a reading habit. Here's a list of some books I've read. Please let me know if you have any book recommendations!
  • I'm a big fan of The Office and Community, and some of my favorite stand up comedians are Nate Bargatze, Gary Gulman, and Sam Morril.
  • I'm also natively trilingual in English, Russian, and Hebrew :)

Contact me

You can reach me at "mpokryva" at "gmail" dot com.