I'm Naman Saini. I turn messy data into reliable pipelines, useful analysis and machine-learning systems — with an emphasis on clarity, evaluation and practical outcomes.
01 / SELECTED WORK
Each project focuses on a problem, a transparent process and a measurable result — not just a list of technologies.
Analysed 9,993 Superstore order lines to identify where discounting erodes profit. Built reusable analysis modules, customer segmentation and an interactive dashboard.
Compared four classifiers across 283,726 transactions and optimised a decision threshold for fraud recall and estimated costs rather than misleading accuracy.
Built a working investigation-assistant MVP that connects retrieval, least-privilege tools, typed agent hand-offs and approval-gated recommendations. Uses synthetic SOC data; not deployed in production.
Cleaned and analysed 18,000 synthetic SOC incidents to explore attack patterns, SLA breaches and response metrics.
Investigated ML and deep learning methods for intrusion detection, focusing on class imbalance, feature engineering and per-attack evaluation.
Prepared approximately 50,000 customer sales records with imputation, deduplication, type conversion and outlier checks.
Looking for more implementation details? Explore all GitHub repositories
02 / WHAT I BRING
Tools matter, but the goal is to make them useful for the people making decisions.
Translate raw datasets into understandable KPIs, dashboards, forecasts and recommendations.
Python · SQL · pandas · Plotly · StreamlitBuild and compare models, address class imbalance and explain performance trade-offs honestly.
scikit-learn · XGBoost · TensorFlow · EvaluationDesign reproducible data preparation, SQL-based systems and early-stage AI applications with auditable outputs.
ETL · MySQL · FastAPI · RAG · MCP03 / BACKGROUND
My background combines computer science, postgraduate AI study and experience supporting Python programming education. I focus on rigorous analysis, reproducible work and explaining technical choices clearly.
Connect on LinkedInUniversity of Bath, UK
Machine learning, deep learning and NLP. Dissertation research in network intrusion detection.University of Bath, UK
Supported Python labs, systematic debugging, code review and accessible explanations of programming concepts.Amity University, Noida, India
Foundations in software engineering, algorithms, object-oriented programming and relational databases.04 / LET'S CONNECT
Interested in opportunities across data analytics, data engineering and applied AI. I'd be glad to discuss where I could contribute.