FATIMUN NISHA
I'm a Data Analyst & ML Engineer from Odisha — I build predictive systems, design BI dashboards, and occasionally publish research on things I actually care about, like making farming smarter.
Who I am
I'm Fatimun Nisha — I did my B.Tech in Computer Science (Data Science & ML) from Centurion University, Odisha, graduating with a CGPA of 8.98. Numbers matter to me, but only when they mean something.
My day-to-day involves writing SQL queries, wrangling data with Python, building dashboards in Power BI, and training ML models end-to-end. I've done this both in a professional setting at Clivertech and in research that got published in two peer-reviewed papers — one of which came out of a project I genuinely loved working on.
What keeps me going is that moment when messy, scattered data finally tells a clear story — one that helps someone make a better call. That's what I'm here for.
Technical Stack
Complex joins, CTEs, window functions — I've spent a fair bit of time writing queries that pull real meaning out of relational databases (MySQL, mostly).
Pandas and NumPy for cleaning and EDA. I like building pipelines that I can actually hand off to someone without a ten-page explanation.
Dashboards that stakeholders actually use — not just pretty charts. XLOOKUP, Pivot Tables, and BI reporting that informs real decisions.
Scikit-learn and TensorFlow for supervised learning, feature engineering, and NLP. I care a lot about model evaluation — accuracy alone never tells the full story.
AWS Certified Cloud Practitioner. I've worked on connecting IoT sensor data into ML pipelines — that project ended up getting published.
Two published papers, and counting. I can write for academic reviewers and explain the same thing to a non-technical audience — both matter.
Where I've worked
What I've built
Built a supervised ML system that recommends the right crop based on soil composition and climate data. Full pipeline — preprocessing, feature engineering, multi-class evaluation. The kind of project where getting the model right actually matters for someone's livelihood.
Took the crop recommendation work further by layering in real-time IoT sensor data. The results outperformed non-IoT baselines by a clear margin — and the research behind it was published in a Routledge edited volume in 2024.
A hybrid recommender that combines collaborative filtering with NLP-based sentiment classification. Multiple models compared, full text preprocessing pipeline. Good reminder that user reviews are data too.
Research
Credentials
Let's connect
I'm actively looking for data analyst and ML engineering roles.
If something I've built caught your eye — or you just want to talk data — reach out.