I'm a Data Scientist

I drive GenAI, product, and stakeholder decisions, across large-scale systems and agentic AI initiatives.

*This is where I'd normally write a long paragraph about my qualifications, but that would take too many LLM tokens.*

I recommend checking out my highlighted projects. They summarize all my experience in quick and fun reads.

Highlighted Projects

LLMs · Graph RAG · Retrieval ⏱️ 9 min read

IBM Agentic Library - Graph RAG for Document Q&A

Built a Graph RAG pipeline with Neo4j and FAISS for structured retrieval. Knowledge-graph modeling improved retrieval precision by 30%+ over naive search.

Explore the Analysis →
NLP · Vision · Detection ⏱️ 7 min read

M(iche)Langelo - Analysis on AI-Generated Art

Built an end-to-end Reddit pipeline for style, caption, and source detection using CLIP, BLIP, and SuSy. Transfer learning to a 3-class SuSy head improved MidJourney and DALL-E detection on real-world art streams.

Explore the Analysis →
Causal Inference · Optimization ⏱️ 12 min suite read

Product Analytics Suite

A combined suite of three analyses covering feature impact, growth allocation, and root-cause diagnosis using PSM, DiD, uplift modeling, and MMM.

Explore the Analysis →

Leadership & Impact

Built analytics products from zero to scale

Founding team member for MyLoanBhai, TestMyPolicy, and FinQy 2.0, establishing product analytics foundations from first release to scaled execution.

Scaled operations across 35 cities and 3 countries

Led data-backed operating workflows across multi-country teams, enabling standardized execution and measurable performance governance.

Lead author on 4+ IEEE papers

Published work spanning predictive maintenance, NLP-to-SQL automation, and ML modeling with 15+ citations across peer-reviewed research.

Global AI hackathon winner

Won first place in a global AI hackathon, ranking 1st out of 750 teams in a 1,311-participant competition spanning 19 countries.

Media Praise

Grand finale collage for Hackathon on Plastic-Free Rivers with AI at REVA University
Featured In ThePrint (ANI PR)

"Hackathon on Plastic-Free Rivers with AI" Grand Finale at REVA University

11 September 2023 · REVA University + Kyndryl

  • 750 teams from 19 countries and 1,311 total participants.
  • Top 15 finalist teams reached the live grand finale demo.
  • Team EcoGuards won 1st place and a prize of INR 1,50,000.
  • Challenge focus: Vision AI for plastic detection, classification, and segmentation from drone imagery.

Read media coverage → · Watch event recap →

Background

Professional Experience

Data Science Researcher (Generative AI and Personalization)

Julius Baer · New York, NY

Jun 2025 – Sep 2025
Gen AI Personalization LangGraph Experimentation
  • Engineered an LLM personalization chatbot using LangGraph and prompt chaining, automating content selection and curation time by 35%.
  • Uncovered a 25% lift in customer content engagement by running causal and exploratory analysis on customer behavior using Python and SQL.
  • Standardized A/B test, guardrails, and reproducibility across cross-functional teams by establishing an experimentation metrics framework.

Data Scientist Consultant (Customer Analytics and Segmentation)

TD Bank · New York, NY

Jan 2025 – May 2025
Customer Analytics Fraud Detection XGBoost + SHAP
  • Validated fraud-detection strategies at scale by executing causal inference and A/B testing on 100M+ transactions in PySpark and Databricks.
  • Elevated fraud model accuracy from 0.20 to 0.67 PR-AUC (235% gain) using an XGBoost model with SHAP explainability for stakeholders.
  • Guided customer segmentation, money laundering reduction, and strategy by assisting with senior management through data-driven insights.

Quantitative Research Apprentice (Acquisition Analytics and Growth Optimization)

Ask2AI · New York, NY

Jan 2025 – May 2025
Acquisition Analytics Growth Optimization MILP
  • Clarified ROI attribution across channels by 22% by deploying uplift modeling and experimentation to isolate incremental acquisition impact.
  • Boosted customer LTV by 12% and reduced churn by 25% by building regression and MILP for acquisition allocation using PyGurobi.
  • Shaped growth strategy and executive decisions by synthesizing forecasting-driven causal insights, lifting acquisition rates across key segments.

Teaching Assistant (Algorithms to Data Science)

Columbia University · New York, NY

Sep 2024 – Jan 2025
Algorithms to Data Science Student Mentorship
  • Mentored 250+ MS students on ML algorithms, analytical reasoning, experiment design, and inference through structured weekly sessions.
  • Designed case studies and assignments on A/B test design, modeling, and data-driven decision-making for applied data science coursework.

Data Scientist Consultant (Predictive Maintenance and Operational Analytics)

Navin Fluorine International Limited · Mumbai, IN

May 2023 – Nov 2023
Predictive Maintenance Operational Analytics
  • Reduced equipment downtime by 35% and cut ingestion latency by 40% by architecting a predictive pipeline with XGBoost and Airflow.
  • Lifted recall 15% and shortened repair cycles by deploying anomaly detection using XGBoost and Isolation Forest to surface high-risk events.

Founding IT Intern (Acquisition and Retention Analytics)

E-Revbay Pvt. Ltd. · Mumbai, IN

Dec 2021 – May 2022
Acquisition Analytics Retention Analytics
  • Increased qualified customer leads by 50% per quarter by leading acquisition analytics pipelines using Equifax data, SQL, and Python.
  • Eliminated 25% of manager intervention time and accelerated response by building real-time dashboards and root-cause pipelines in Tableau.

Artificial Intelligence Intern

Verzeo

Jun 2021 – Aug 2021
Data Preprocessing Computer Vision CNN
  • Processed tabular diamond datasets for classification, improving model accuracy by 17% through data cleaning and feature engineering.
  • Built flower image recognition with CNN techniques, achieving 85% accuracy while reducing error rates by 15%.

Education

M.S. Data Science

Columbia University · New York, NY

Sep 2024 – Dec 2025
GPA 3.7/4.0
  • Relevant Courses: Agentic AI; Fintech and Data Economy (PhD elective); Big Data Analytics; Statistics; Applied Machine Learning

B.Tech Honors (Computer Engineering, Data Science/Analytics)

NMIMS University · Mumbai, India

Jun 2020 – May 2024
GPA 3.9/4.0
  • Focus Areas: Computer Engineering, Data Science, and Analytics

Research Publications

Predictive maintenance for metro systems

Google Scholar · Lead Author

Demonstrated a sensor-driven pipeline that predicts equipment failures so operations teams can schedule targeted maintenance and reduce downtime.

Diabetes detection optimized for recall

Google Scholar · Lead Author

Prioritized recall in model design to reduce missed diagnoses, improving early detection reliability for clinical use.

NLP to SQL for mobile learning

Google Scholar

Built an NLP-to-SQL interface that lets non-technical users query student and CSV data directly from mobile devices.

Semi-supervised disease prediction

Google Scholar

Applied semi-supervised methods to leverage unlabeled clinical data and improve prediction robustness in ambiguous diagnostic cases.

Contact

Always happy to discuss data science, experimentation, causal inference, and GenAI applications.

Email: abhitayshinde@gmail.com LinkedIn GitHub