My Projects

This is the website you are browsing right now! My personal website describes who I am in a nutshell, and has links to my calendar, projects and ways to contact me. I built this website from scratch, learning more about web development in the process.

View the site: www.stevencamina.com

View the source on Github.

• Date: August 2017
• Type: Web Development

Fanny Friendly LLC is a limited liability company registered in the State of Wisconsin that promotes and sells bidet sets to retail and wholesale markets. The company was co-founded by my uncle, so I built this website for him from scratch and maintain it pro-bono.

View the site at www.fannyfriendlyllc.com

View the source on Github.

• Date: August 2017
• Type: Web Development

TaxServPro provides tax preparation and consultation services to clients across New York, New Jersey and Connecticut. My uncle runs this business from his home in New Jersey.

View the site at www.taxservpro.com

View the source on Github.

• Date: March 2018
• Type: Web Development

I starred as Doug in my friend's (Jorge) 15 minute short film titled "The Bug". Doug is a young computer programmer who tries to get the girl (Faith) while battling his code - and himself.

This was my first time acting in front of a camera, and I had lots of fun doing it :)

Here's the IMDB page for the film - link.

My friend uploaded the film to Vimeo. The site is password protected. Shoot me a message if you want access.

• Date: 2010
• Type: Acting

For one of my last MuleSoft road shows (aka Summits), I put together a demo that was still being discussed even after I left the company. In the demo (which I run through live on stage), I take a picture of an object with my phone and upload it to Instagram. Mule, running on CloudHub, would then grab the picture and calculate its average color. It will then pass that color to an MQTT queue (a popular IoT protocol), which will then be pulled by another Mule running on a local Raspberry Pi. The Raspberry Pi is networked to a set of Philips Hue bulbs, and the color of those bulbs would change to the color pulled by the Mule on the Raspberry Pi.

It sounds complicated, but it's quite simple. At a high level, there's a Raspberry Pi set up (always a crowd pleaser), and when I take a picture with my phone, the lights around the room suddenly change color. Add some showmanship, and this demo is a guaranteed crowd-pleaser.

We decided to update the demo in MuleSoft HQ, this time using Twitter instead of Instagram. In the updated demo, each time someone tweets mentioning @mulesoft, a set of lights (shaped in the MuleSoft logo) brightens. Here's a link to the blog post describing the demo at HQ:

https://blogs.mulesoft.com/dev/anypoint-platform-dev/demo-iot-lights-raspberry-pi/

View the source code on Github.

• Date: February 2015
• Type: Technical Demo

This is my Master's research at MIT. Here, I attempted to model research landscapes (e.g. "renewable energy") through the automatic generation of hierarchical structures (taxonomies) comprised of terms related to the landscape's name (e.g. "renewable energy"). I compared several different taxonomy generation algorithms, testing each using data gathered from research papers in publication databases like Scopus. I compiled metadata from the papers, used the terms found therein to create a taxonomy of related terms, and evaluated the resulting taxonomy's validity.

The taxonomy generation algorithms I considered were Dijsktra-Jamik-Prim's (DJP) algorithm, Kruskal's algorithm, Edmond's algorithm, Heymann algorithm, and the Genetic algorithm.

Read my thesis here: http://dspace.mit.edu/handle/1721.1/62632

• Date: September 2010
• Type: Master's Degree Thesis

This is the first O'Reilly book I co-authored with MemSQL colleagues. The book talks about real-time data architectures and the different technologies that are needed to build a real-time data pipeline.

O'Reilly website link: http://www.oreilly.com/data/free/building-real-time-data-pipelines.csp

MemSQL website link: http://www1.memsql.com/oreilly.html

• Date: June 2015
• Type: Authored Book

This is the second O'Reilly book I co-authored with MemSQL colleagues. The book talks about building a real-time data pipeline to ingest data into a central database, then creating real-time dashboards, and applications for machine learning and predictive analytics on top of that.

O'Reilly website link: http://www.oreilly.com/data/free/the-path-to-predictive-analytics-and-machine-learning.csp

MemSQL website link: http://www1.memsql.com/oreilly-predictive-analytics.html

• Date: June 2016
• Type: Authored Book

While I was at MemSQL, I was surprised to see that there wasn't any collateral that answered the question "What is MemSQL?" apart from verbose technical documentation. With that in mind, I decided to put together a couple of videos with the help of our contract videographer and upload those to YouTube.

I received amazing reviews for the first video of this series. It quickly crossed the 1k viewers threshold, and many of the MemSQL new hires who started after the video was released specifically reached out to tell me how helpful this video was during their interview process.

Video 1: https://www.youtube.com/watch?v=HoompXCG5Mk

Video 2: https://www.youtube.com/watch?v=CDkmbq8EggI

• Date: November 2016
• Type: Video Tutorial

This is the talk I gave at the 2017 Spark Summit East conference. The talk was only 15 minutes long, but it covered a lot of information.

Video link: https://www.youtube.com/watch?v=gapTJYxyV2E

• Date: February 2017
• Type: Conference Talk

In this webcast, I first gave an overview of machine learning and data science, then went deeper into the components of a machine learning pipeline which I called "experimentation" and "productization". I then described at a high level what technologies are needed to build a machine learning stack and gave a few detailed examples of the "productization" component as seen in industry. I wrapped up with a demo and some Q&A.

Webcast Recording: http://www.oreilly.com/pub/e/3855

Slides: https://www.slideshare.net/MemSQL/building-the-ideal-stack-for-machine-learning

• Date: January 2017
• Type: Webcast

While I was at MuleSoft, I delivered many talks around the US, South America, and Europe. I also recorded several webcasts, which featured a technology talk, demo and some Q&A. I was always very hands-on (i.e. product demos) in my presentations. Here are links to the webcasts I delivered while I was at MuleSoft:

Mule 101

Mule 201

Connect Faster with Mule 3.5

SOA and Legacy Modernization

• Dates: 2011-2013
• Type: Webcast

While I was at Oracle, developers of Portal (Oracle BRM) needed to integrate TimesTen with their product. The current monitoring mechanism was to use a TimesTen utility called 'ttStats', but that was insufficient for their needs. With that in mind, I worked with one of the TimesTen QA engineers (Doug) to build something better. Doug wrote the backend that can query a TimesTen database and output XML. I then created a UI wrapper around his tool. I used Flex to build the frontend and quickly iterated on it. My app polled a specific URL at a given interval and displayed the output onscreen. The user of my app could then drill down on the output.

The app was proprietary and used only internally. I don't have the source code, so I posted a couple of screenshots here.

• Date: 2012
• Type: Application Development

For Oracle OpenWorld 2012, I built an application that was used in the Intel keynote talk in the main hall. Thousands were in attendance for the talk given by Diane Bryant, VP & GM of the Datacenter & Connected Systems Group at Intel. Her talk featured my demo, which showed TimesTen Database performance at scale - i.e. with more hardware, throughput scaled linearly while response time remained fairly constant - running on Intel's new SSD hardware.

I built the front-end and part of the back-end of the application. Code was written in Java.

Video of the demo: https://www.youtube.com/watch?v=N4Uvbm8JhDs

• Date: September 2012
• Type: Application Development