Honeywell Life Sciences
Software engineering · 2025
Analyst. Developer. Curious by default.
I’m Karan. I turn complex data into clear decisions, and build the tools that make them useful.
Business analytics meets software engineering.
From the question to a working product.
A few places.
A broader perspective.
Industry, education & research
Tata Motors
Data & process automation · 2023
Alliance Manchester Business School
Business analytics · Operational research · Risk
VIT Vellore
Computer science & engineering · 9.1 / 10
Fujitsu Research of Europe
Low-hallucination AI · Business risk reporting
Different settings.
The same curiosity.
Selected work
Products & decision systems
OneCelerity
Marketing intelligence, connected.
A marketing-intelligence SaaS bringing Shopify and Meta Ads into one view of profit, campaign performance, forecasts and supplier decisions. Built with FastAPI, PostgreSQL and an AI analyst with structured outputs and fallback logic.
Orbit LMS
A calmer way to run a school.
A Flutter and Firebase learning platform connecting role-based access, assessments, fees, report cards and analytics. One product for the people on every side of a school.
StaffSyncScheduling model
StaffSync
Better decisions for A&E operations.
Hospital analytics combining cost, fairness and skill-aware scheduling with statistical sampling, audit trails and XGBoost breach prediction.
International CoalOptimisation model
International Coal
The trade-offs behind a cleaner fuel mix.
A power-plant decision model evaluating fuel mix, carbon pricing, SO₂ limits, biomass and emissions-control investment across operational and financial scenarios.
A little context
The person behind the workI like understanding how things work.
Then making them work better.
My background sits between computer science and business analytics. I’m interested in the whole problem: the quality of the data, the trade-offs in a decision, and the software people use to act on it.
I’ve worked on engineering and automation at Honeywell Life Sciences and Tata Motors, and built products of my own. Across all of it, the aim is the same: make complex systems more useful and easier to trust.
Away from a screen: football, Borussia Dortmund, and music.
Python & SQLFastAPI & PostgreSQLOptimisation & MCDAStatistics & forecastingLLMs & NLPFlutter & Firebase
Industry experience
Engineering with a business outcomeHoneywell Life Sciences
Software Engineer Intern
Developed an AI-assisted compliance-scanning workflow for pharmaceutical quality reviews, using NLP, LLM suggestions and time-series analysis through FastAPI.
Automated 20+ UI/API regression tests, saving approximately 10 hours per QA cycle and increasing test coverage by around 10%.
Jan — Jul 2025Bengaluru, India
Tata Motors
Software Engineer Intern
Replaced an outsourced vehicle-testbed conversion workflow with an in-house ETL tool, standardising raw data into Q-DAS-compatible DFQ records.
Processed approximately 2,300 files weekly, reducing batch turnaround from three hours to about two minutes and avoiding around ₹7.8 lakh in monthly vendor costs.
Aug — Oct 2023Pantnagar, India
Research & exploration
Making the answer more dependableResearch of EuropeOngoingMSc industry research collaboration
AI decision support.
With less room for hallucination.
Building a Low-Hallucination LLM Decision Support System for Business Risk Reporting.
Exploring cybersecurity risk and incident reporting, with an emphasis on traceable evidence, structured outputs and dependable narratives for business decisions.
About the research partnerPublished research · IJIRT · May 2024Sentiment-Based Rating Model
Sentiment-Based Rating Model
using VADER and TextBlob
A large-data project using Coursera course reviews: cleaning, tokenisation and standardisation, followed by sentiment-to-rating models evaluated with accuracy, RMSE and weighted F1.
Read the publicationFoundations
Education & continued learning2025 — 2026
The University of Manchester
MSc Business Analytics:
Operational Research and Risk Analysis
Alliance Manchester Business School
Linear programming · Optimisation · Applied statistics · Forecasting · Business intelligence
2021 — 2025
Vellore Institute of Technology
B.Tech Computer Science & Engineering
CGPA 9.1 / 10
Software engineering · Data systems · Machine learning · Applied mathematics
R.A.N. Public School CBSE Class XII · 97.8% · 2021
Additional learningAWS Cloud Foundations · AWS Cloud Architecting
Internet of Things · Applied statistics
