Platform: Backend Infrastructure
Languages: Golang, Java
Creation: Summer 2017
Details:
Worked on two projects for improving the developer experience for mobile engineers at Uber, while also optimizing existing services. The first project revolved around creating a Go service to provide remote device connection for application testing of the mobile applications. Utilizing third party services to provide remote device connection, the device labs project provides an API and device agnostic wrapper to various functionality with devices including connection and disconnection. It was usable through a friendly commandline interface, with logging and metrics collection.
The second project build an event pipeline for differential builds from Phabricator and Jenkins. It provided strong metrics and a notification system capable of failure commenting.

Platform: Backend Web Development
Languages: Ruby (Ruby on Rails)
Creation: Present
Details:
Co-Founder and CTO of Pomily.

Position: President
Date: Present
Details:
President of DevX at UCLA.

Platform: Web
Languages: Javascript (MKRN)
Creation: Present
Details:
I am currently working as an organizer on the development team for LAHacks - the annual 36 hour hackathon at UCLA, working on building the foundations of an entire system infrastructure for the future of LAHacks. My work primarily focuses on the backend development for the current state of the site, to enable people to create accounts and apply for the hackathon from March 31st to April 2nd of 2017. We focused on using the MKRN (Mongo, Koa, React, Node) stack. My major contribution was on building the user services, controllers and endpoints focusing on resetting passwords and admin privileges, along with some of the tests. As I continue to work with the rest of the team, I intend on making huge contributions towards more of the backend along with the future projects and ideas we have for the system.
The site and the application can be found on the website.

Platform: iOS
Languages: Swift
Creation: Present
Details:
I am currently on the board for Creative Labs as the iOS Project Lead. In summary, Creative Labs serves as the bridge between designers and developers at UCLA to bring the two disciplines together to build amazing quarter long projects, and give anyone and everyone the opportunity to learn the skills to become either a developer or designer through our weekly workshops and speaker series.
As the iOS Project Lead, I have worked with leading a beginners iOS workshop in the Fall of 2016, where I helped 7 students understand the fundamentals of Swift and iOS development through a simple Todo application. This Winter, I intend on leading an advanced iOS project by recruiting about developers and designers to work on a iOS project from the ground up to release on the App Store.
More information can be found on our website.

Platform: iOS
Languages: Swift, Objective-C
Creation: SD Hacks, October 2015
Awards: Best Use of IBM Bluemix [IBM at SDHacks], Most Entrepreneurially Viable Hack [Perkins Coie at SDHacks], Hack That Best Connects Us [ViaSat at SDHacks], Most Innovative Award [Facebook at ACM Hacker Expo 2015], Individuals Impacted [IBM at Bluemixathon]
Details:
Future without borders was conceived and created around the peak of the Syrian Refugee Crisis in order to try and provide help to the fleeing refugees that entered these new countries, hopeful to start a new life.
Following a pattern similar to that of AirBnb, we decided to build out an iOS application that would be ideally used by the citizens of the countries accepting refugees at the time, to allow these "hosts" to see any nearby refugee family that might be in need of assistance. The app would allow the host to provide aid to the family selected by picking them up from the nearby locaiton and hosting them for a few days to help them start their new lives.
On the refugee side, we assumed the access to a smartphone would be limited, and so we used Twilio's API to provide a unique number that the refugees could text to register them as requiring assistance. Based on cellular location we would provide said location to the iOS application.
This project won three sponsored awards at SD Hacks and continued to win two more awards from IBM and Facebook. Unfortunately, due to several unfortunate circumstances, the project fell on the back burner and to this day remains a MVP. More information can be found on the project's Devpost.

Platform: iOS
Languages: Swift, Objective-C
Creation: HackSC, November 2015
Awards: Best Mobile Hack
Details:
The weekend of HackSC was also the weekend of the horrible Paris Attacks. Noticing, the lack of an efficient method to warn people, family, friends and law enforcement, we decided to create an easy, single tap solution. From that came Lighthouse, a single tap SOS button located on your home screen. The version built at HackSC used Parse as the backend and was able to notify family members and friends listed in the app of your danger.
After seeing the usefulness of this application, we continued to build it out over a few months, making the application more robust and marketable. We created our backend in Ruby on Rails, hosted on Heroku. The iOS application evolved to have its own networking interface, included requested tracking, pulled daily updated crime events in the area, and acted as a miniature social media application, aimed at bringing together people for the purpose of safety and crowd-sourced danger reports.
After noticing that we couldn't find a decent monetization strategy, the application was abandoned. More information about the application can be found on the Devpost and on the landing page.

Platform: iOS
Languages: Swift, Objective-C
Creation: LA Hacks, April 2016
Awards: Third Place
Details:
After reading a paper on Convolutional Neural Networks and Art we decided to create a mobile application that would make use of the torch implementation to allow everyday users to get a feel for the true power of Artificial Intelligence in their hands.
After running some tests, we found that on a computer such as a Macbook Pro Retina 13" the image rendering took several minutes to a couple of hours, we decided to move the computation to Amazon EC2 servers with more powerful graphics cards. The iOS application was written to take in two images either through your camera roll or through an image address and would send the images to the server managed by node.js, where a BASH script would take the combined image and send it back to the iOS application which would display the results. The app ran well within 5 minutes per image, and was one of the very first versions of mobile CNN's. It placed third overall at LA Hacks.
The application was slated to be worked on during Summer of 2016, to create scheduling mechanisms for server loading, but another app beat us to market. More information about the application can be found at the Devpost.

Platform: Nexsys-3 FPGA Board
Languages: Verilog
Creation: CSM152A - UCLA, May 2016
Details:
As the final project and submission to CSM152A, my teammate and I built Doodle Fall. This game is a simple one, where the objective of you, the tiny green doodle, is to stay alive as long as possible, by falling downards along the blocks that appear, avoiding hittint the top and bottom of the screen. The score was determined as a function of time. Similarly, the difficulty of the game was also a function of time, wherein the speed of the falling would increase as you lasted longer in the game. The green blocks were powerups that would slow down the speed of falling.
This game was made entirely in Verilog, using an FPGA, Joystick, VGA, Seven Segment LED Display and buttons. The game would store the highest score, and would allow you to reset the game. The joystick allowed you to move horizontally with the speed of movement as a function of the movement of the joystick. The VGA took our code and created a simplistic view on any computer monitor. The video above demonstrates the key features of the game.

Platform: iOS
Languages: Swift
Creation: CSM117 - UCLA, May 2016
Details:
In order to demostrate our knowledge of networking and real-world application in CSM117, myself and my teammate Connor built a small application that leveraged Firebase as our backend.
Using Facebook's iOS SDK for login and user authentication and Firebase for data storage and retrieval, we built out an application wherein you, the user can go about your daily life, and if you find anything intersting in the location you're currently at, you can leave either a message about the situation or the place within that location on a global map. Any other user of the app can find that message/photo and view it.
The aim of the application was to allow users to share their experiences in a more intimate manner.

Platform: Oculus, Myo Armband, Muse Headband
Languages: C#, Python
Creation: CalHacks 2.0, October 2015
Awards: Top Five at CalHacks 2.0
Details:
In an attempt to work with hardware, my teammates and I took it upon ourselves to make neurology a more immersive experience. The result was BrainSpace VR, wherein you dawn an Oculus Rift headset to find yourself viewing the 3-D model of a human brain. Equipped with the accompanying Myo Armband, you as the user are able to click on different parts of the brain to learn more about it, and are able to rotate and zoom into the brain to experience it like never before.
Omar Ozgur was primarily involved with writing some of the C# code to work with the Oculus, while I was involved with using Python and the BeauitfulSoup Library to scrape various webpages to find the different informaiton about the different parts of the brain. Together we worked to additionally include the Muse Headband, to allow simultaneous viewing of your own brain wave pattern. More about the product can be found on our Devpost.

Platform: CryEngine
Languages: C++, Lua, Cryengine Flowgraphing
Creation: CS188 - UCLA, June 2016
Details:
Using CryEngine my teammates and I created an MVP game with Artificial Intelligence mechanisms. The premise of the game was that the player was a man sent back through time to steal the very first iPhone model in order to throw themselves into fame and fortune. The game was built with only a single level to demonstrate its potential. The player is thrown into an environment where they are expected to get through the level either by alerting no guards, or by taking out all the guards in their way to get to the final destination. The score is determined by the speed and efficacy with which the player reached the goal.
We programmed the guards to have varying levels of AI using CryEngine's in-built AI mechanisms with slight tweaks of our own, such as giving the guards a proximity sensor based on noise, visibility and distance of the player to the guard. So the guard would move from standby, to suspicious to alerted mode. The player would be able to hide in a group of employees to detract attention.
The game had the ability to be AI controlled based on player performance metrics. Using a weighted function based on the time, targets killed and other factors, we programmed the AI to complete the level achieving the best score possible. Thus the game would be able to switch from user mode, to AI mode. The video above describes in great detail each of our contributions and the results. A brief report with further links can be found here.

Product: AutoCAD
Platform: Visual Studio
Languages: C++
Period: June 2016 - September 2016
Details:
Worked to implement an effective and platform independent serialization mechanism, leveraging JSON, to create an Inter Process Communication channel. This project was built as a workable unit test integrated into AutoCAD with code written to be readily extensible. The serialization involved taking low level calls and packaging them into a custom built messaging schema. Throughout the internship I gained experience in discussing progress and workarounds with professional senior software engineers and software architects. I learnt how to navigate a huge code base consisting of million lines of code , while learning the conventions for building industrial strength projects. This greatly improved my ability to modularize and componentize projects.