Contact information

Summary

I design, develop, maintain and run software systems with a focus on long-term maintainability, and lead small teams doing so.

What do I do?

Software design

Software systems need both functionality and properties such as flexibility, adaptability and ease of operations. I know how to isolate layers, invert dependencies, partition systems, evict state, add immutability, split interfaces and make tons of other small and large decisions to make it happen.

Requirements change all the time. I know how to do an MVP to avoid spending months implementing features nobody wants.

Complexity creeps in. I know how to prune unwanted statefulness, remove layering violations, keep the interfaces minimal.

All software is distributed now: Web SPA and mobile apps are distributed systems too. I know how to manage upgrades, do a staged rollout client- and server-side, collect metrics and log messages, how to deal with old versions of client software suddenly reappearing out of the blue and how to version REST protocols to avoid breakages while staying flexible.

Software development

Give me a task, I’ll do it. Or better: give me a problem and I’ll solve it.

I always try to see the big picture: can this problem be solved without writing new code? By deleting existing code? By a change in an internal process? Or by using an external service (price, SLA)? Does this problem need to be solved at all?

(Distributed) team leadership

I have experience leading small engineering teams (from 2 to 8 people), both on-site and distributed (with timezones spread from UTC-4 to UTC+8), both in commercial setting and in a voluntary cooperation around an open source project.

I like this kind of work and would like to do it again, especially in an environment where management and technical leadership are separate roles.

Reengineering and refactoring

Is your legacy system a mess? Does implementing a new feature take ages, do bugfixes cause more problems than they solve, do system administrators have to apply enormous amount of effort to keep it running?

This is something I am able to do and like to tackle. I have experience decluttering various software systems, getting rid of creeping statefulness and streamlining operations.

Tools

I know my tools. I have introduced CI into several projects, bootstrapped an embedded Linux distribution and all of its infrastructure, and wrote a countless number of tests, makefiles and scripts.

DevOps / SRE

Server configuration management is so XX century (I’m saying this as a major contributor to CFEngine).

Read-only stateless containers are a good thing (I added a container-related tool to util-linux).

I’d like to work on a project that either already uses or switches to existing tools, such as Docker and Kubernetes. Implementing infrastructure bits and pieces from scratch is a lot of fun and one can spend enormous amount of time on it, while it does not bring any customer-facing value.

What I can not (will not) do

  • NEW FEATURES NO MATTER WHAT-NEW FEATURES NO MATTER WHAT-NEW FEATURES NO MATTER WHAT. There are many people who are way too happy to create new features and never look back at the mess created. I’d rather take additional time making sure the new functionality integrates with the rest of the system and does not contribute to a Big Ball of Mud.
  • Client-side all day long. I’m reasonably proficient in working with client-side functional UI libraries (React, Reagent), but it is not my forté.
  • PHP.
  • Perl.
  • Windows as a mandated development environment.

What do I want?

  • Remote work (unless the company offices are located on Malta).
  • Flexible work schedule, to a reasonable extent.

Projects/Companies

Open Source

I have contributed numerous patches to various open source projects, mostly fixing problems I have encountered.

Are you running Ubuntu or Debian? Say apt-cache show libssh2-1 and see the Maintainer field.

Are you running Linux? Check out the AUTHOR stanza in man 1 unshare.

Hola! / Luminati (2013+)

Wearing my SRE hat, I have improved insight into performance of Hola services, by adding

  • centralised application-level logging, and
  • application-level metrics

that made it much easier to pinpoint and resolve performance problems and bugs.

I have created several internal-facing statistic dashboards for Hola VPN and also user-facing statistics for Hola Spark CDN.

While working on Luminati I have implemented a custom in-memory DB to track the status of 1M+ peers of Luminati network that decreased the price of hardware running this database by 8x compared to the off-the-shelf DB.

Debian (2006+)

I maintain several packages in Debian, mostly “scratching my own itch”.

My Debian experience usually makes me a “resident Debian/Ubuntu expert” in companies I work for.

CFEngine (2011-2013)

I have thrust CFEngine from SunOS 4 times into 21 century (not single-handedly, but pushing hard for the changes), converting it from R&D-ware to a reliable software package along the way, adding unit and system testing, continuous integration, discernible release management and other useful software engineering practices.

OpenInkpot (2007-2012)

A Linux distribution for eBook readers. It was fun while it lasted until Kindle killed it.

I have started it, have built the infrastructure (from bugtracking to cross-compiling Debian packages in CI environment to repository manager) and was BDFLing it until it withered out.

Also I’ve got the experience managing distributed team in addition to the previous experince working in one.

Parallels (2004-2007)

For Plesk for Unix I have created (in a team) a tool for transferring contents of competing hosting panels to Plesk, which was then repurposed as a backup system.

OpenFusion project was my first spell as a software architect. I have designed SSO mechanism for Parallels apps, and also a Web applications packaging format (Wikipedia page).

Other

From To Name What it was/is about?
2009 2010 IPlinux Unwillingly started a Debian-based embedded distro to build OpenInkpot on top of something after previous base distribution got abandoned by the primary developer. Worked OK, but OpenInkpot was always the main focus.
2007 2009 ALT Linux I was maintaining distro bugtracker, bringing development documentation into order and creating internal web apps for analysis of development process.
2003 2004 Axmor As a member of IBM Solutions Group team I was doing R&D work for IBM, evaluating new cool technologies of the time for potential usefulness for IBM projects. Also doubled as system administrator.
2003 2003 Novosoft An intern! My first “real” place of work. Billing system based on Cisco NetFlow.

Keywords/Experience

30Y+ experience in

Russian language (native).

10Y+ experience in

C, Linux, Make, Debian, Korn-like shells, POSIX/SUSv3.

5Y+ experience in

C++, Javascript, Node.js, TCP/IP, Python, git, CFEngine.

Norwegian language (upper intermediate).

3Y+ experience in

MongoDB, various Internet protocols, REST interfaces, ElasticSearch, AWS.

1Y+ experience in

Clojure, Perl, various Unices, XSLT, Java, Django, Eclipse RCP, SQLite, PostgreSQL.

Have a good grasp of

OO, functional, structured and other kinds of programming. OO and functional design. Various tools and toolchains (CI, builds, testing etc).

Toyed with

Various Lisps, Go, Docker, Haskell, React, Ruby, OCaml, Rust, Erlang, C#, J2EE (old one), Google Cloud Platform, Kubernetes, Ansible, Salt, Puppet, and everything else on the front page of Hacker News.