Open to Lead / Principal Data Science roles
I’m Mohit Bhatnagar, a Senior Data Scientist with 10+ years across retail, loyalty, marketing and industrial operations. I build models stakeholders can trust and ship them as tools teams actually use.
AuthorAI Engineering: From First Principles to Production
Everything a hiring team needs before a first call — role, depth, domains and what I’m looking for next.
Showing 10 projects
Recommends the products and categories most likely to succeed in advertising campaigns, using historical performance and machine-learning signals.
~40% improvement in campaign success
An explainable ensemble model that grades promotional campaigns as good, bad or neutral, replacing a manual review process.
80%+ less manual campaign analysis
Tracks behavioural shifts across promotional phases with segmentation, reporting and decision-oriented measurement.
Informed customer-retention strategy
Combines artificial neural networks, Random Forest and XGBoost to predict cricket match outcomes.
85%+ prediction accuracy
Unsupervised learning on plant data to identify efficiency regimes and improve boiler operations in manufacturing plants.
₹0.72 Cr annual savings
Recency, frequency and monetary segmentation combined with lifetime-value modelling for personalised targeting.
25%+ campaign ROI improvement
Cleans unreliable influencer-campaign data, forecasts sales with uncertainty and serves scenarios through Streamlit and FastAPI.
Live demo, CI and model card
Topic in, narrated educational video out: local planning, image and video generation, offline text-to-speech and a Rust compositor.
100% local, end-to-end pipeline
Turns original story prompts into structured episodes, panel plans and rendered comic pages, with safety and IP guardrails.
Safety-aware, multi-backend generation
Turns plain-English questions into guarded, read-only MySQL queries and returns a summary, result table and download.
Safe natural-language access to data
Professional projects are summarised at a non-confidential level; figures are as measured on the client engagement.
Long-term professional delivery, backed by technologies you can inspect in the public repositories.
The Layered Learning Edition — From Everyday Intuition to Research-Grade Systems.
View on AmazonWork, stack, public evidence and this site’s privacy model.
Recommendation, classification, forecasting, customer and retail analytics, industrial optimisation, marketing prediction, local generative AI, media pipelines and natural-language access to business databases.
Python, PySpark, SQL, XGBoost, CatBoost, SHAP, MLflow, FastAPI, Streamlit, Gemini, LangChain, Ollama, MySQL, Rust, FFmpeg, ComfyUI, Power BI and Tableau.
Professional projects are described only at a non-confidential summary level. Four public GitHub projects have full source code and case studies; private repositories and unreleased products stay private.
No. There are no forms, accounts or analytics cookies. Contact happens through direct email, LinkedIn and WhatsApp links.
Roles, consulting or collaboration — email is the fastest way to reach me.