Aditya Shirodkar

Data Engineer

About Me

Hello! I’m Aditya Shirodkar, an experienced data engineer based in Bangalore, with a strong background in building and maintaining ETL pipelines, data analytics, and machine learning. I also have an interest in generative AI and its applications.

Bio

adityashirodkar2020@gmail.com

Phone

+91 8806460040

Address

Bangalore, India

Work Experience

Data Engineer at Culture Shift

Manchester, UK
Sep. 2021 - June 2023

Culture Shift is a SaaS startup whose main product is an incident reporting app with an in-built analytics engine.

Built and maintained numerous ETL pipelines that synchronised the company’s product (single-page React app) and database (Postgres) with various vendor tools like Mixpanel, HubSpot, and Google Drive.

Developed custom data pipelines to analyse client site performance, product feature usage, and company’s own key performance indicators (KPIs).

End-to-end development of a data request system for company use, allowing colleagues to access data from within the app itself.

Statistical anomaly detection, broadcasting such incidents as spikes in reporting rates to Slack for further investigation.

Conducted an R&D project using machine learning techniques to audit the structure of the incident reporting forms and estimate the optimal form structure to attain higher form completion rates.

My data was instrumental in securing contracts with the UK Cabinet Office, prestigious Russell Group universities, and multi-national companies such as EY.

Test-driven development, comprehensive dev. database, type-annotation and linter use, detailed code review, Agile workflow.

Tools: Python, TypeScript (node.js), AWS Lambda/S3/EC2, Gitlab, Docker, SQLAlchemy, Poetry

Wind Resource Analyst Intern at the National Institute of Wind Energy

Chennai, India
July, 2017 - December 2017

“Practice school” project with BITS Goa. Part of the Wind Resource Assessment team at NIWE.

Identified potential sites for wind farms using NIWE-monitored data and GIS data in the state of Tamil Nadu, meeting government energy policy criteria and NIWE client requirements.

Data Analyst Intern at HDFC Bank Head Office

Mumbai, India
Jan, 2017 - June, 2017

“Practice school” project with BITS Goa. Part of the X-sell team of the Credit Risk Analytics division at HDFC Bank.

Introduced parameters to account for the urbanity of business loan customers, for better X-selling in rural areas with fewer spending pressures. This improved the accuracy of existing models to calculate the risk of default for business loan customers.

Mined data from textual point-of-sale transaction details to create labels for company use.

Education

Masters in Data Science from Lancaster University

2020 - 2021

Distinction

Specialism in Computing.

AI Project: “Solvers for the infinitely-many armed bandit problem.” Expanded upon eleven existing algorithms for the traditional finitely-armed case.

NLP project: Categorised words and sentences as English, Czech, or Igbo with 99.9%+ accuracy.

Estimated property prices in England and Wales for GBG using geographical data. Achieved street-level confidence.

Created a distributed file sharing network (Hadoop-based) to handle billions of rows of Google server data in real time using on-campus Linux terminals.

Thesis project: “Predicting in-patient mortality” with NHS UHMB. In-patient mortality was to be predicted, generalisable for trust data on a per-patient level. Utilised undersampling techniques to offset the size (millions of rows) and highly imbalanced nature of the dataset; achieved high recall results at the cost of precision, considered acceptable as one is "better safe than sorry" in the medical sector.

Masters in Renewable and Sustainable Energy Technologies from Northumbria University

2018 - 2020

Distinction

Modelled sizing, power flow, penetration, reliability, and LCOE of a hybrid renewable energy system for a term project.

Dissertation: “Thermal analysis of a substrate embedded with two different microencapsulated phase change materials.” Utilised Salome-meca/Code_Aster (Python-based simulation software/FEA solver) to show that embedding two types of microencapsulated PCMs of different latent heats into a substrate can provide better thermal storage properties than using either one when distributed in the right proportion.

BE Mechanical Engg. and MSc Biological Sciences (double major) from the Birla Institute of Technology and Science

2012 - 2017

First class with honours

Minor in Philosophy, Economics, and Politics

"Celestia" astronomy club: organised Storming Aurora, an astrophysics conference (2012).

"Searock" semi-professional band competition organiser (2012 - 2016).

Debate and quiz enthusiast with the Literary and Debating Club (2012 - 2016).

Professional Skills

Python

Experienced

Classic data science: pandas, scikit-learn, etc.

Experienced

Data visualisation: matplotlib, seaborn, ggplot2 (R)

Experienced

Deep learning: PyTorch

Experienced

TypeScript (node.js)

Novice

Docker

Novice

Git

Experienced

AWS Lambda, EC2, S3, etc.

Novice

RMarkdown, LaTeX

Experienced

Projects

Python > Generative AI

Article research tool

Allows you to investigate articles provided as their URLs using an LLM. Can be run in a browser using streamlit.

Python > Generative AI

Restaurant idea generator

Creates a restaurant name and menu items from the provided cuisine. Can be run in a browser using streamlit.

Python > Generative AI

Image to speech story

Creates a speech story from a submitted image. Can be run in a browser using streamlit.

Python > Deep learning

PyTorch ML tools

Helpful implementations of basic neural networks, building up towards generative AI.

Python

Interactive tutorial on basic finance concepts

An interactive text-based tutorial to teach some basic finance concepts to children aged 8+.

Contact

+91 8806460040

adityashirodkar2020@gmail.com