Portfolio — 2026

Soumyadeep
Roy

Data Scientist · Machine Learning Researcher · MSc Computer Science. Focused on building practical data and machine learning solutions across predictive modelling, applied AI, and analytics.

About meView projects
Python ✦Machine Learning ✦RDKit ✦Cheminformatics ✦PyTorch ✦scikit-learn ✦SQL ✦AWS ✦Python ✦Machine Learning ✦RDKit ✦Cheminformatics ✦PyTorch ✦scikit-learn ✦SQL ✦AWS ✦
01

About me

I'm a data and machine learning professional with an MSc (Merit) in Computer Science from the University of Leeds. My experience includes AI/LLM data validation at Innodata and ongoing machine learning research at the Indian Institute of Science, where I develop predictive models for chemical reaction yields using molecular representations, transfer learning, and ensemble methods.

My work spans applied machine learning, data analysis, cheminformatics, and cloud-based workflows—from predicting dengue fever outbreaks and detecting fraudulent transactions to modelling complex chemical reactions.

I enjoy turning difficult real-world datasets into reliable, interpretable analytical solutions and am especially interested in applied AI, scientific machine learning, and data-driven research.

LocationBengaluru, India
FocusML · Data Science · Research
LanguagesEnglish · Hindi · Bengali
EducationMSc CS — Leeds (Merit)
MobilityOpen to relocate
Contact+91 98322 05144
02

Work Experience

Jul 2026 – Present
ML · Research
Machine Learning Research Intern
Indian Institute of Science (IISc) · Bengaluru, Karnataka · Research Internship
  • Developing machine learning models for chemical reaction yield prediction using molecular representations and experimental reaction data.
  • Built and evaluated regression pipelines using Random Forest, Extra Trees, pretrained neural models, and class-specific modelling approaches.
  • Applied transfer learning with learned reaction embeddings and conducted hyperparameter optimisation through cross-validation.
  • Improved minor-product yield prediction beyond the reference benchmark while achieving near-benchmark performance for major-product reactions.
  • Working with Python, RDKit, PyTorch, scikit-learn, Pandas, NumPy, and Jupyter Notebook.
Apr – Jul 2026
AI · LLM
Analyst — AI/LLM Practice
Innodata Inc. · Remote · Fixed-term Contract
  • Investigated and validated structured datasets supporting generative AI and large language model pipelines.
  • Detected annotation inconsistencies and data inaccuracies through systematic review, SQL checks, and Python-based validation.
  • Automated repetitive validation tasks with Python, improving efficiency across high-volume review workflows.
  • Documented QA findings, recurring anomaly patterns, and issue resolutions for research leads and cross-functional teams.
Jun 2024 – Jul 2025
Data · Operations
Sales Advisor — Technical Support & Data Operations
Sky UK · Leeds, United Kingdom · Full-time
  • Analysed customer service records and operational data to identify recurring technical and service disruption patterns.
  • Investigated system and service records to support root-cause analysis and faster issue resolution.
  • Maintained structured documentation within CRM and reporting systems, supporting accuracy and auditability.
  • Communicated investigation findings to technical and operational teams to support process improvement.
May – Aug 2022
Cloud · AWS
Cloud & Data Engineering Intern
Kritter Software Technology · Bengaluru, Karnataka · Internship
  • Assisted development of cloud-based data workflows and automated ETL pipelines using AWS S3 and Lambda.
  • Performed data validation, cleaning, and transformation for backend processing workflows.
  • Supported monitoring and debugging of automated pipelines to improve data integrity and reliability.
Dec 2019 – Jan 2020
Python · Backend
Backend Developer Intern
Workex · Bengaluru, Karnataka · Internship
  • Assisted Python backend development, debugging, unit testing, and QA validation.
  • Maintained version-controlled code and documented technical observations for senior developers.
03

Selected Projects

001
Research Project · Indian Institute of Science
Machine Learning for Chemical Reaction Yield Prediction

Ongoing research focused on predicting chemical reaction yields from molecular and experimental information. Developed a class-aware modelling pipeline combining learned reaction embeddings with tuned ensemble regressors, and compared baseline, pretrained, and classifier–regressor approaches through repeated cross-validation.

Minor-product RMSE: 5.34, improving on the 6.48 reference benchmark. Major-product RMSE: 8.53, close to the 8.40 benchmark.

PythonRDKitPyTorchExtra TreesRandom ForestTransfer LearningCheminformatics
002
MSc Research · University of Leeds
DengAI — Dengue Spread Prediction

End-to-end machine learning research predicting dengue fever spread across San Juan and Iquitos using epidemiological and environmental time-series data. Evaluated regression, Random Forest, SVR, and PCA-based approaches for weekly case prediction.

PythonPandasscikit-learnRandom ForestPCATime Series
View on GitHub ↗
003
Independent Project
Credit Card Fraud Detection System

Built a full machine learning pipeline for detecting fraudulent transactions in a highly imbalanced dataset. Developed preprocessing, model training, evaluation, inference, and a FastAPI prediction endpoint deployed online.

PythonPandasRandom ForestFastAPIREST APIDeployment
View on GitHub ↗
004
Internship Project · Workex
Recruitment Chatbot Workflow

Collaborated on a conversational chatbot for a recruitment platform. The workflow automated candidate screening across conditional decision nodes covering education, language fluency, experience, and salary expectations, with automated follow-up scheduling.

PythonChatbot DesignJSONWorkflow Automation
04

Skills & Tools

Programming
  • Python — Pandas, NumPy
  • SQL — MySQL, PostgreSQL
  • C++
  • FastAPI & REST APIs
Machine Learning
  • Classification & Regression
  • Random Forest & Extra Trees
  • Feature Engineering
  • Cross-validation & Tuning
Scientific ML
  • RDKit & Molecular Data
  • Reaction Yield Prediction
  • Transfer Learning
  • Learned Embeddings
Deep Learning
  • PyTorch
  • TensorFlow fundamentals
  • Neural representations
  • Model evaluation
Cloud & Engineering
  • AWS S3, EC2 & Lambda
  • Git & GitHub
  • Jupyter Notebook
  • Data pipelines
Analytics
  • Exploratory Data Analysis
  • Statistical Modelling
  • Excel, Power BI & Tableau
  • Matplotlib
05

Education

Master of Science
Advanced Computer Science
University of Leeds, United Kingdom
Sep 2022 – Sep 2023 · Merit

Modules included Machine Learning, Data Science, Cloud Computing, Algorithms, Scientific Computation, and Bio-Inspired Computing. Dissertation: DengAI — Predicting Disease Spread Using Machine Learning.

Bachelor of Technology
Computer Science & Engineering
SRM Institute of Science & Technology
Jul 2017 – Jun 2021 · First-Class Distinction

Core coursework included data structures, algorithms, database systems, software engineering, and object-oriented programming.

06

Certifications

Data Analytics with Python

NPTEL · IIT Roorkee

95%

Machine Learning

Stanford University · Coursera

Building Web Applications in PHP

University of Michigan · Coursera

Blockchain & Decentralised AI Workshop

Microsoft Student Partners Club

Let's work
together.

Open to opportunities in data science, machine learning, research, and data analytics—particularly roles where applied AI meets real-world scientific or business problems.

Emailroymayo593@gmail.comLinkedInSoumyadeep Roy ↗GitHubgithub.com/roymayo593 ↗Phone+91 98322 05144