Data Scientist • GenAI • MMM • MLOps

Hi, I’m Ankit Wahane.

I build end‑to‑end ML systems that move business metrics — currently focused on GenAI agents, MMM, and production‑grade MLOps. Based in Bengaluru, open to roles in India, EU, and UAE.

Currently @ TransUnion • ZS Associates • IISc alum
A H Wahane
Currently working on AI Agents for Insight Generation at TransUnion

Education

IISc
MTech in Communication and Networks
Indian Institute of Science, Bengaluru
VIT Pune
BTech in Electronics & Telecommunication
VIT, Pune

Work Experience

Data Science Consultant @ TransUnion, Bengaluru (Oct 2024 – Present)

  • AI/ML in marketing analytics; building MMM models & simulators.
  • GenAI with LLM Agents to automate/simulate scenarios incl. QBR automation.

Associate Consultant @ ZS Associates, Bengaluru (Jul 2022 – Jul 2024)

Drug Shortage Prediction

Transformer-based forecasting combining internal & news signals; identified shortages 1–2 months earlier.

Supply chain Transformer
Pharmacy Orders Forecasting

XGBoost + ALS for retention/cross-sell; E2E deployment on AWS; added OpenAI explanations.

Pharmacy AWS
Bayesian MMM & Budget Optimizer

Evaluated marketing impact; optimized spend; GPT summaries & scenario comparisons. Est. +20–25% revenue uplift.

MMM Bayes

Data Science Associate @ ZS Associates, Bengaluru (Aug 2020 – Jun 2022)

Predictive Lead Scoring Model

Deployed on SageMaker; engineered features; improved conversions & targeting.

Lead RF
Gate‑Checked Bags Forecasting

CatBoost + ANN; 10–15% reduction in under‑prediction using flight & pax features.

Bag CatBoost
Insights Extraction Pipeline

NLI (PubMedBERT) on Databricks to power SPN insight mining.

SPN1 SPN2

Machine Learning Intern @ Apra Labs, Bengaluru (May 2019 – Jul 2019)

Gaze capture iOS app for data collection; gaze estimation model; iOS game & driver safety demo.

Engineer (Automotive BU) @ Tata Elxsi, Bengaluru (Aug 2016 – May 2017)

RTOS operations; coordinated system services & IPC for audio amplifiers.

Technical Skills

ML & DL

Supervised/Unsupervised, Transformers, BERT, GPT/LLMs, CNNs

GenAI & Agents

RAG, LangChain/LangGraph, Multi‑Agent systems, Prompting & Eval

MLOps & Data

AWS, GCP, Databricks, Docker, CI/CD, PySpark

Modeling & Stats

Bayesian Stats, Optimization, Time Series, MMM

Programming

Python, SQL, Git

Libraries

PyTorch, TensorFlow/Keras, scikit‑learn, XGBoost, JAX, Pandas, Plotly/Matplotlib/Seaborn

Articles

SafeTensors vs Pickle vs GGUF: Secure & Efficient Model Formats

Published

A deep dive into three popular model serialization formats — performance, security, compatibility, and when to use which in production. Includes conversion code recipes and a quick decision guide.

Read More →

Personal Projects

GenAI • MMM • Forecasting
QBR Agents

QBR GenAI Agents Orchestrator

Engagement, gatherer, and synthesis agents generate editable PPT QBRs. LangGraph + Streamlit.

LangGraphRAGPPTX
MMM

Bayesian MMM + Budget Optimizer

Hierarchical MMM with response curves; JAX‑accelerated inference; ROI‑based spend optimizer.

PyMCJAXPlotly
Multi‑Agent Hub

Data Scientist Learning Hub (Multi‑Agents)

CrewAI Researcher/Writer/Reviewer agents to generate curated study notes.

CrewAILLMsStreamlit
Drug Shortage

Drug Shortage Prediction

Transformer‑based early warning using internal + news signals; identified shortages 1–2 months earlier.

TransformersXGBoostLLMs
Alter Igo

Alter Igo – GenAI Code Assistant

Generate‑fix‑execute analytics loop with charts & tables.

AgentsRAGEvaluation

Awards

Know Me Beyond Work

Travel • Photography • Cooking
photo1 photo2 photo5 photo7 photo11 photo6 photo9 photo10

Contact

Open to roles and collaborations. Email me or drop a message — I usually reply within a day.