Nareg Amirian Megan

Internship

Software Engineer Intern

I'm currently interning at Platform9, a company that creates, manages, and scales cloud infrastructure. I'm Working on the Platform9 Managed Kubernetes team integrating Helm 3 into the PMK service.

Course Staff

Teaching EECS

CS 61A Lab Assistant: Answered questions during lab and ofﬁce hours pertaining to Python, Scheme, SQL, or any other course topics. Resolved technical issues with both the students and the course including homework and project submission errors.

CS 61C Tutor: Teach groups of students about C/low-level code, assembly code, basic microprocessor design, computer memory, and data/thread-level parallelism. Help students in labs, discussions, and office hours.

Computer Architecture Research

Gemmini Project for ML hardware accelerators

Coming Soon.

Machine Learning Developer

ML Developer @ Cal Launchpad

I've worked on various projects in Launchpad involving classical ML, deep learning, and RL. As a part of education committee I helped lead workshops on different ML topics. I'm currently working on the external committee to expand the professional and educational reach of our club.

Multi-Robot Cooperation

Research Assistant

I worked in a group of undergraduate researchers within the Biomimetic Millisystems Lab on multi-robot cooperation. We used tools such as VREP, CAD, and physical robots to test and improve path planning, cooperation, and task completion abilities of a group of connected robots.

Data Analytics Intern

Data Analytics in Astrophysics

The Wisconsin IceCube Particle Astrophysics Center studies neutrino data from the IceCube research facility in the South Pole. I Used Python and SQL to analyze neutrino data and develop datasets of commonly misclassiﬁed neutrino events from the IceCube neutrino detector. I Leveraged a platform called Zooniverse for human analysis of data which improved the quality of the training data. Improved neutrino classiﬁcation algorithms SVMs, neural nets using curated data.

Startup Internship

Software Engineer Intern, Brokerage Startup

While working for a startup aiming to streamline the real estate brokerage process I helped build the data intake pipeline using image/text processing and optical character recognition OpenCV, regex, etc. I utilized Google Cloud APIs such as Google Vision and Gmail. Assisted in the design of the API used to integrate backend functionality with ReactJS frontend. Helped create tests and integrate the platform for product release.

Verilog Generator for HFT

Generating hardware in Verilog for basic quant trading strategies

High Frequency Trading and quantitative trading relies largely on the speed of decision making and execution. FPGAs can be used to create task specific hardware and replicate that hardware thousands of times over, allowing for highly efficient and parallelized computation. In this project I wrote a Verilog generator. When given a json file which describes mathematical rules for executing trades the generator creates both the hardware description and a test bench which can be used to virtually run the hardware using Icarus.

Java ANN Library

Artificial Neural Network framework in Java

In order to learn the inner-workings of neural networks I wrote my own basic ANN library in Java including my own matrix processing library. In the end I was able to get the network to accurately model basic functions such as xor.

Visual Q/A System

Launchpad Project: Visual Q&A

The goal of this project was to recreate and improve upon state of the art VQA model architectures. The goal of VQA models is to answer a question about an image (see above). We tried many different approaches ranging from K-nearest neighbors to our own experimental architecture. Our final model used a recurrent form of attention attempting to emulate how a human would comprehend a sentence as they read it. For more specifics see the README on our github.

Emergent Behavior in RL

Launchpad Project: Emergent Behavior in Reinforcement Learning

Emergent behavior in reinforcment learning is the behavior of agents in an environment that have formed due to the relationship it has with other entities in its environment. In our case we studied emergent behavior in symmetric game environments like pong and snake. You can see an example of a form of emergent behavior in the gif above, as one agent learns to rapidly slide the paddle to hit the ball at a speed the other player can't respond to fast enough.

Ancestral Tracing (Moorjani Lab)

Launchpad Project: Ancestral Tracing in collaboration with Moorjani Lab

Inferring ancestry from genome data of reference populations from two (or more) reference populations admixed together. Used a hidden markov model through the implementation of the viterbi algorithm and SVMs.

Web Dev

Web Dev

I've had experience developing the frontend and backend of different sites. Most of my frontend work has been with vanilla HTML/CSS/JS, and my backend experience is primarily with Python and PHP. One main example of my frontend ability is this website, which I designed and built from scratch (uses bootstrap, particles.js, and isotope). One backend project I've worked on recently is a COVID-19 contact tracer using the LAMP stack.

Nareg Amirian Megan

Portfolio

Coursework

Skills

About Me