I was studying for AWS Certification and I created a study platform from my own study notes.

Safety Compliance Software for Transportation Carriers.Using ML we process documents and compile them into reports.Anything that can automate and improve Saftey Compliance we do it.

I'm the seasonal CTO for PrepAnywhere. During the school year, I'm focused on PrepAnywhere.PrepAnywhere is* Video solutions for high school math * An LMS used for in-person tutoring all done iPads

Monsterbox was my web-dev firm. I use the company now for when I do tech consulting. I had multiple remote developers and at any given time we could have 8 projects on the go.I was an "Interim CTO" for multiple early stage startups.The focus was rapidly building MVPs and providing support for spillover work for other design/dev firmsMonsterbox is now just my consultancy company as I no longer have an interest in building MVPs for startups.

When I was the CTO of OneClass the core business model was the reselling of university study notes. We crowdsourced study notes by incentivizing students to upload theirs in exchange for gift cards.I joined OneClass when it was pre-seed, there were multiple competitors in the space, the company was being DDoS'd and it was presumed that during the start of the semester and at exam time it would be DDoS'd again (and we guessed right). On top of that, the codebase was cobbled together by remote developers and required an entire rewrite which had to be completed in 3 weeks before the start of the new school year.I mitigated the first DDoS attack by spreading the load across multiple servers when we didn't have the money to pay for an expensive DDoS mitigation service. AWS Shield and CloudFlare were non-existent or ineffectual for the style and size of DDoS attacks at the time and googling resulted in no DIY solutions.BackboneJS came out the same week my first week at OneClass and I took the gamble to adopt it in the rewrite on top of everything else because I felt that javascript-frameworks were the going to be the future and this gamble paid off giving us a competitive edge over our competitors in terms of user experience.After the initial growing pains, I spent most of my time focusing on growth hacking, fraud detection and managing thousands of PDFs being uploaded daily.

Brigle was a social aggregator. The technical achievements were impressive but the product failed to monetize which I think was due to the timing of product. You could attempt to compare the service for HootSuite. This was before the age of web-frameworks, so I had to roll my own proto web-framework. The challenges were web scraping and polling APIs from multiple social services when standardization wasn't a thing. We needed a reliable queueing system and so I adopted RabbitMQ. OAuth was an emerging technology and we made use of it where we could. To reduce the initial onboarding of new developers I used Vagrant and Chef which were new at the time to get create VM development environments.I think if this product was a few years later when web-frameworks were out and when we saw more interest in companies trying to leverage social media it would have been of great success.

Teambox was an open-source project manager that was a direct competitor to Basecamp.When I joined, we started over from scratch heavily borrowing from my existing codebases Markadee and TeacherSeat and developed what would be known as Teambox Version 2.Since it was an open-source project it required strict processes especially towards testing. At the time BDD, Cucumber and Capybara were emerging testing concepts and I adopted these technologies early on at Teambox and wrote the bulk of the user-acceptance features and specs which if you look at the archived codebase that early effort remained well in-place well after my departure. We used at the time EngineYard for our host and TextMate was the popular IDE of choice.We had no web-framework since the concept was non-existent at the time. I had chosen to use PrototypeJS over JQuery because PrototypeJS had powerful ruby-like utility functions similar to underscore/lodash but despite its technical superiority over JQuery, JQuery won the "javascript library wars" due to community popularity and so in retrospect I think we should have instead used JQuery.It sounds trivial now but the two most difficult challenges were multi-file upload and real-time updates.Multi-file upload was tricky for both client and server side because it was not well documented at the time. I remember we had achieved multi-file upload before Basecamp hoping it would give us a competitive edge only for them to release multi-file upload a few weeks later.To solve real-time updates I coded in Python Twisted. NodeJs had just come out this year and Web-Sockets was still in working draft. When I had left the company this I believe was being moved over to NodeJS and Faye.Also, Resque (background jobs via Redis) had just come out so I migrated away from delayed_jobs where we saw improved performance. We used NewRelic as our APM when it didn't cost an arm and a leg.