Close

Amy Pajak


Machine Learning Engineer



View CV

About Me

Amy



Hi! I'm Amy :) I work as a Machine Learning Engineer at Meta.
I graduated with First-Class Honours in BSc Artificial Intelligence and Computer Science from the University of Birmingham in 2021. My undergraduate thesis focused on comparing and applying statistical and deep learning techniques for dimensionality reduction in medical imaging.

I enjoy collaborating with researchers and engineers to bring novel ML solutions to fruition, experimenting with models and ML techniques as well as staying on top of the latest research.

Industry Experience

Meta

Machine Learning Engineer

Goldman Sachs

Machine Learning Engineer

Associate in AI Research team within Data Science and Machine Learning group.

  • Experimented & selected appropriate state-of-the-art ML models to solve real-world problems and apply to tasks within NLP systems.
  • Optimised, fine-tuned and implemented deep learning models on information extraction tasks such as NER, Summarization and Question Answering to improve accuracy and efficiency.
  • Researched and identified new ML methods through independent study, experimentation and participation in internal knowledge-sharing communities.
  • Implemented and productionised models and algorithms to improve the performance of existing systems, processes, and products.

IBM

Software Developer

Placement year as L3 Engineer in cloud and cognitive software primarily working in C++.

  • Developed and implemented new features and efficiency improvements to allow increased serviceability.
  • Completed Triages and code fixes (APARs) to resolve production-level issues.
  • Achieved speed improvement of over 60x through implementing asynchronous threading to achieve batch sending of messages. Added to production with gMock and component tests.

Hertzian

Software Developer

Full-stack developer at AI start-up building back-end systems for data analysis and front-end web applications to display outputs for a range of customers. Primarily worked in Python and JavaScript.

  • Front-end development using JavaScript/JQuery, Ajax, Bootstrap, HTML and CSS.
  • Worked with Python and JSON for data analysis and used frameworks such as Flask and Django for output.
  • Managed an NHS project called ‘Hertzian Health’ which analyses patient feedback data from various wards across the U.K. and produces informative outputs on ward performance.

Education

University of Birmingham

Sept 2017 - July 2021

BSc Artificial Intelligence and Computer Science Hons (with Industrial Year)

UCL (University College London)

Sept 2020

Quantum Technologies Summer School

Truro College

June 2017

A-Levels

Personal Projects

AI to Detect and Track Infectious Disease Outbreaks

I explore how Bayesian networks can be utilised to create an intelligent system that effectively performs biosurveillance of infectious disease outbreaks. By combining the power of machine learning with modern issues involved in safeguarding public health, I share how achieving early, reliable detection can assist medical professionals and government bodies to provide time-critical response and treatment.

View Poster

Marketing Emotion Response Detector

This project aims to classify the emotion on a person's face when viewing advertisement media into one of seven categories, using deep convolutional neural networks. The model is trained on the FER-2013 dataset which consists of 35887 grayscale, 48x48 sized face images with seven emotions - angry, disgusted, fearful, happy, neutral, sad and surprised.

View Project

Lego Mindstorms EV3 Robot

Using computer vision the EV3 robot will navigate obstacle courses clearly without collisions. Given a location as a goal to reach the robot will detect obstructions in its path and safely clear them by determining a better route.

Hackathon Projects

Algothon 2019

Using sentiment analysis and social media data this hack predicts stock directionality. In the modern world it's no surprise that the sheer volume of activity and sentiment online can massively affect consumer-facing businesses. Using forward monthly predictions to overcome random short-term noise we trained Random Forests models to generate alpha with predictions on a month-long trading horizon.

View Project

MusicLab

HackTheMidlands 2018
MusicLab is an AI melody generator. Based on a short input (e.g. typing in notes, singing into a microphone or playing an instrument) MusicLab will create a melody from songs it's been trained on using Markov chains.
Who said you need to learn how to compose music to actually compose music!?

View Project

News.ai

A.I.Camp Hackathon 2018
Inspired by the rising popularity of the phrase "Fake news" we created a news article bias checker. Both an independent website and a browser add-on, News.ai analyses the article you're reading and judges its bias based on a number of factors such as language, semantics, sources, author and more.

View Project

Skills

Get in Touch