Logo

Hi, my name is

Shourya Sharma

I'm a Machine Learning Engineer.

Machine learning engineer with 4+ years of experience building scalable data pipelines and backend architecture. Language agnostic with Python expertise and proficiency in Java and .NET Core. Experienced in designing event‐driven systems that integrate ML models for high throughput production environments.

About

Hey I'm Shourya, nice to meet you! Please take a look around.

I love to code and create applications that are user-friendly and improve the lives of those around me. I have 3+ years of experience in software development, and graduated from the University of Surrey with a first class honours BSc in Computer Science. I currently specialise as a backend developer, but am looking to expand my skillset to incorporate data science and machine learning as well.

Skills

These are the technologies I've worked with

Github

Github

Java

Java

Node

Node

React

React

Python

Python

PostgreSQL

PostgreSQL

Atlassian

Atlassian

Docker

Docker

Projects

Check out some of my recent projects below:

activity-detection

The Suspicious Activity Detection System is a Python application that monitors a live video stream from a camera, detects suspicious activit...

CollabCode

Collaborative code editor developed as part of the Discover Intern Hackathon. Allows for live editing from multiple users.

100-days-of-ml

Collection of notebooks / code / datasets used during my 100 Days of Machine Learning Coding, as shown in https://github.com/Avik-Jain/100-D...

authentication-framework

Custom Authentication Project written in dotnet and react that handles email sign up, verification, and password resetting, all while using ...

ChessAI

A chess bot trained using a dataset containing 5 million positions that were converted to binary for classification in terms of win / loss. ...

NLP-Cuisine-Chatbot

A chatbot application developed with the DistilBERT transformer model for named entity recognition and a RoBERTa transformer model for inten...

Work

  1. ComplyAdvantage

    MACHINE LEARNING ENGINEER

    • Currently working on the design and implementation of a next-generation adverse media ingestion system, leveraging Convolutional Neural Networks (CNNs) and LLMs to effectively filter and classify over 100 million articles monthly.
    • Contributed to the development of an innovative similarity filter mechanism by testing clustering performance with Locality-Sensitive Hashing (LSH) and Facebook AI Similarity Search (FAISS), ultimately resulting in an 80% reduction in duplicate filters for adverse media articles and significantly improving processing efficiency.

  2. Qube Research & Technologies

    QUANT TECHNOLOGIST

    • Integrated several new python data sources through which traders could implement strategies and generate revenue. Decreased ETL of data source onboarding by 2 days by developing an in-house automated integration testing library in .NET, which was critical to increasing the velocity of implementing new trading strategies in production.

  3. ComplyAdvantage

    SOFTWARE ENGINEER

    • Developed and maintained an orchestration layer of Django, FastAPI and Flask microservices responsible for monitoring and running 1000s of web scrapers on international sanction sources - this data was the core of our product, supporting millions of dollars of revenue and was expected to be available to tight SLAs.
    • Architected a solution for migrating data from a storage-based fetch system of 40 million records from a mongoDB database to a data-streaming model built on Apache Kafka through Kafka Connect and confluent-python - this allowed data to be readily available to clients in near-real time.
    • Implemented an audit dashboard connecting the orchestration layer to ElasticSearch to gather analytics on data changes within sanction sources, a significant feature that was required by enterprise-level clients and is now frequently used for auditing purposes.

  4. Discover Financial Services

    BUSINESS TECHNOLOGY INTERN

    • Developed an internal dashboard used to manage and monitor all software engineering projects in the payments pipeline, using a tech stack consisting of Spring Boot, Flask, React and Redux, PostgreSQL and Apollo GraphQL. This was used by all engineering teams within the company to accelerate delivery.
    • Created a standardised testing framework for all projects using React.js by documenting and presenting the usage of Jest and Cypress, improving our software reliability and code coverage by around 30% on the dashboard application.
    • Developed an automated CI/CD process to integrate said testing framework into Jenkins, allowing for the application to be thoroughly tested before being pushed into a production environment - this improved software reliability.

  5. BAE Systems

    SOFTWARE ENGINEER

    • Developed a Python and React self-hosted web application responsible for utilising live ship and radar data as parameters for a genetic route- planning algorithm. This was demo'd to royal navy commanders and scheduled for test deployment on live ships.
    • Mentored software engineering apprentices on the team with one-to-one sessions focused on introducing best practices such as usage of git for source code management, and the importance of updating and maintaining documentation.

  6. VS Design Solutions

    FULL STACK DEVELOPER

    • Worked as a part time full stack developer using React.js and Ruby on Rails, developing a ticketing system similar to Jira which was used by employees to track progress of projects being undertaken at the company. This improved the visibility of all projects within the company, and thus improved communication with clients extensively.
    • Developed and maintained the company website in Ruby on Rails and managed day-to-day dashboard updates on the company intranet.