+3517599940408
ABOUT
I'm a fourth-year undergraduate studying Electrical and Electronic Engineering at Imperial College London, but I am currently doing a year abroad at École Polytechnique in Paris, France. In my third year, I took courses mostly in mathematics, machine learning, communications and control. I am currently following the “data science” track at École Polytechnique which focuses mostly on machine learning, natural language processing and big data systems. I'm fluent in English and Portuguese and speak French to an advanced level. I spent my last summer as an intern on the Quantitative Investment Strategies Trading desk at Barclays (2019).
With a strong background and diverse skill set, I’m passionate to keep working on more projects and expand my knowledge and experience, below are only a few projects I've worked on. Keep exploring my site to learn more information about me and reach out directly with any questions.
PORTFOLIO
Projects I've worked on
PIUS- CO-FOUNDER
June 2018- present
Pius scrapes the Christian Bible and returns verses for any given keyword or prayer as an input. Users can then share verses, join discussion forums or follow their favourite churches.
As part of a team of 3, I was responsible for creating the back-end for the mobile application. I built a fast search engine using spaCy, NLTK and Textrank in Python and SQL. Using these tools I leveraged natural language processing and text tagging for similarity analysis and keyword extraction. The back end was then deployed on GAE (google app engine) and connected to our app (coded in react native) which is now available on the app store.
DEEP LEARNING PROJECT
January 2019 - March 2019
Coursework with the objective of generating descriptors for noise-corrupted local image patches. With reference to the baseline, an improved model was proposed. It maintains a modular architecture but changes both the de-noising and descriptor networks. The improved de-noising block uses the full UNet with Adam on MSE loss, the descriptor model uses L2-Net with Adam on triplet loss with an anchor swap.
SECOND YEAR FINAL PROJECT- BIKESAFE
January 2018 - June 2018
Built a functioning anti-theft GPS bike tracker with a fully integrated mobile application (using Firebase authentication).
The developed prototype is installed in the seat tube, and uses a combination of GPS and accelerometer data to detect when the bike is being stolen. GSM is used to communicate with the users’ phone, and an app overlays the location data from the bicycle onto a map (image above).
PCB DESIGN - SWITCHBOARD
January 2018 - June 2018
Designed a simple PCB using Altium Circuit maker. 4 touch switches were designed to manually determine the frequency of a Schmitt trigger oscillator. This was done via an MBED module and an OLED display.
FPGA- VERILOG DESIGN
September 2017 - January 2018
I designed an echo synthesizer with variable delay using a 2-port RAM delay block. Using this module I then created a voice corrupter by changing the pitch of the speaker using cross fading of two separately delayed speech signals.
ARM ASSEMBLER PROJECT
September 2017 - January 2018
As part of an optional Christmas project I created a multiple precision multiplier using Karatsuba's algorithm. This is then used to implement a fast subroutine for a reciprocal and square root functions using Newton Raphson methods.
FIRST YEAR FINAL PROJECT- EEROVER
January 2017 - June 2017
In a team of 7, we designed and constructed a wireless moon rover with capabilities of detecting: IR, magnetic, radio and ultrasound signals. We used an Arduino board as the intelligence unit and the Blynk app as the controller for our rover.
INTERNSHIPS & PROFESSIONAL EXPERIENCE
March 2019 - October 2019
BARCLAYS QUANTITATIVE INVESTMENT STRATEGIES - INTERN
Over the course of 6 months, I worked on 2 projects as part of the Quantitative Investment Strategies Trading desk.
Project 1: I worked on the creation and analysis of volatility target indices. Decomposition into greeks and calculating P&L using Monte Carlo and finite difference methods with a large emphasis on an accurate method of reproducing implied volatility.
Project 2: I built a proof of concept for a client-facing website for creating and optimizing portfolios comprised of Barclays indices from scratch. This was built using Dash and Flask in python as well as CSS. The project was then extended into decomposing portfolios by asset class, sectors and single stocks as well as conducting macro factor analysis on a variety of indices. I also created and tested a REST API to segregate the front-end and back-end functional libraries.
Interested in learning more about my professional background?