I'm available for hire as a consultant. Here are some of the things I can do for your company:
Teach your team about consistency. At Shyp the biggest oncall problem we had was in driver assignment; two drivers would get assigned the same pickup, or a pickup would get "assigned" but have no driver, and a driver would get stuck. I wrote a state machine and a transaction library to fix these. I can teach your team the same techniques to write faster, more correct database queries.
Productionize your culture. You shipped a bad error last week; how can you minimize the likelihood of shipping another one next week? Does the oncall rotation feel like an overwhelming burden for your team? The answer goes deeper than "focus more when writing the code", into a lot of human factors areas. I implemented a lot of tools and procedures at Shyp that helped the team learn from its mistakes, ship reliable code, and brought the number of pager incidents to an extremely low value. I'll work with your team to figure out where the problem areas are, and recommend/implement solutions.
Rewrite it in Go. Do you have a server that no one understands, that needs better throughput, or fails in unpredictable ways? Is your team interested in kicking the tires with Go for a new project? I'm a contributor to the Go programming language, I've written (and rewritten) several services in Go for several companies, and can do the same for you. This will include a testing and deployment strategy, a method for loading configuration from the environment, and for starting and testing a local server.
Find problems in the build. Intermittent build failures train your team that build failures are okay, and then you might miss a more important error that comes through. Or, an intermittent build failure might represent a real problem. I can teach your team how to isolate and force an intermittent problem to repeat on demand, then figure out how to fix it.
Speed up the build. When I started at Shyp it took up to 100 seconds to boot up the database for a single test. That's down to about 3-5 seconds. I can apply the same techniques to your team's build, production codebase, or framework. I can also teach your team how to write faster tests.
Find performance problems in production. Let's figure out why that page loads so slowly. Here's a sample of what I can do - I found a 10-20% speedup in Go's JSON encoding library recently.
Run design/content strategy reviews for your documentation. I maintained the documentation for the Twilio API from 2011 to 2014. I ran user tests to figure out where people got stuck when figuring out how to use Twilio. We used a lot of the lessons from that to design great documentation. I talk more about this in "How to Write Documentation for Users that Don't Read" at Write the Docs 2015.
Run design/strategy/correctness reviews for client libraries. I maintained Twilio's helper libraries from 2012 to 2014 and have written many API clients; I gave a talk about designing great client libraries at Twiliocon in 2013. I can work with your team on API and client library design.
Other Stuff. This is not an exclusive list! I've learned a ton on the fly, and can probably do that with your company as well.
A typical contract is for 1 month of work. I am normally booked 1 to 3 months in advance.
I'm open to shorter contracts, but prefer longer ones. For a shorter contract, we'll divide my monthly rate by the number of days you want, and add a small amount of overhead.
Please note that work will be conducted in accordance with my ethics policy.
Built a billing pipeline to help Segment account for its AWS costs. Built an admin dashboard with fine-grained permissions for viewing data about accounts.
"Kevin came in for a week and went over our entire backend. He found a bunch of security holes, helped us set up better ops practices, and figured out ways to improve our database performance. In the end, everything was running smoother and more secure. We call him the one-man S.W.A.T. team for SF startups, and already booked his next available slot 👍" - Ivan Zhao, Co-founder
Built Logrole, an extremely fast, fine-grained Twilio log viewer, in a one month engagement.
Feature development and improving the testing/local development infrastructure.
Massively improved the team's velocity and ability to identify and fix errors, and deploy them to production. Rewrote two unmaintained C servers in Go. Transitioned server-side software stack from physical machines to AWS.
Code reviewed, identified, fixed, and shipped features for a Node.JS app serving the Otto API. Added several libraries to improve team velocity, ease testing and standardize the codebase.