Headline

Data Scientist @ Epsilon | VIT ’24

About

Two engineers ship the same model. One writes "developed ML pipeline." The other tells you what changed because of it — I'm the second kind. I'm a Data Scientist at Epsilon, where I build production ML and GenAI systems for CPG and healthcare clients — causal inference and campaign measurement, heterogeneous uplift modeling (Double-ML / R-learner CATE estimation), and multi-agent RAG architectures. My daily stack: Python, PySpark, Databricks, and AWS. The foundation came early. I graduated with a B.Tech in Computer Science (Data Science specialization) from VIT with a 9.05/10 CGP. Along the way I authored 2 peer-reviewed publications (ICMLDE 2023 / Elsevier and IEEE Engineering Informatics 2023) and picked up certifications in Databricks, AWS, and Adobe Analytics. What I care about most: models that survive contact with production. Measurement that stands up to scrutiny. GenAI systems that do real work instead of demos. Always happy to talk multi-agent systems, causal ML, or GenAI in production — message me here on LinkedIn.

Location

Tumkur

AI stack

anthropic apis

claude code

claude

langchain

langgraph

Microsoft Copilot

Work experience

Data Scientist 1 at Epsilon, Aug 2024 - Present · 2 yrs 4 mos, Hybrid, Built a heterogeneous uplift modeling pipeline in PySpark/Databricks using Double-ML/R-learner CATE estimation on a ~2M matched-panel CPG campaign, enabling audience-level targeting decisions instead of single average-lift reads Designed and shipped a multi-agent GenAI insights system (LangGraph/LangChain) with router, text-to-SQL, RAG, and QA/verifier agents — letting non-technical brand stakeholders query campaign data in plain English, with every numeric claim verified before delivery Diagnosed and resolved production ML reliability issues including non-deterministic CATE outputs, label-encoding faults, and null-imputation bugs, stabilizing pipeline outputs for repeatable client measurement Delivered causal campaign-measurement and propensity-modeling analyses for CPG and healthcare clients on AWS/Databricks, translating model outputs into targeting and budget recommendations

Data Analyst at Epsilon, Feb 2024 - Aug 2024 · 7 mos, On-site, Analyzed large-scale campaign datasets using PySpark and SQL on Databricks, producing measurement readouts and audience profiling reports for CPG clients Built decision-tree segmentation and lift-based labeling workflows that fed downstream uplift models — work that contributed to promotion to Data Scientist within 7 months

Computer Science Intern at Hindustan Aeronautics Limited, May 2023 - Jun 2023 · 2 mos, Bengaluru, Karnataka, India, Developed a Flight Phase Detection system that detects the Phase of the Flight based on the sensor data.

Senior Core member at Kannada Literary Association, Sep 2022 - Apr 2023 · 8 mos, Vellore, Tamil Nadu, India

Education

Bachelor of Technology - BTech, Computer Science Engineering with Specialization in Data Science, Vellore Institute of Technology

Secondary High School, Mathematics and Computer Science, Deeksha Centre for learning

Public links

github: https://github.com/makamamogh23-arch

leetcode: https://leetcode.com/Makamamogh

portfolio-zeta-amber-69.vercel.app: https://portfolio-zeta-amber-69.vercel.app/