Manish Halwale

+•• ••••• •••••••••••@•••••.com

Lead Assistant Manager, Data Science & Analytics · EXL Services Pvt. Ltd. · 3 yrs

I approach data science and analytics with a highly logical and pragmatic mindset, driven by a deep sense of accountability and determination. My diligence ensures that I consistently deliver precise, impactful solutions, operating with the autonomy needed to architect robust analytical frameworks.

  • Open to Bangalore · Hyderabad · Mumbai
  • Hybrid, On-site, Remote
  • In 60 days

Ask about Manish

Skills

Claimed

KerasOCR UsagePyOCR UsageOSLA Technologies UsageClient FocusAnalytical ThinkingTrade-off AnalysisDefining ConstraintsKPI AnalysisRandom Forest ModelingLogistic Regression ModelingRegression AnalysisCatBoost Modeling

Proof of Experience

Lead Assistant Manager, Data Science & Analytics

EXL Services Pvt. Ltd. · Dec 2023 – Present

  • Deliver end-to-end data science solutions for global banking clients, architecting forecasting and predictive models across 95+ KPIs and 2,600+ branches, achieving under 3% MAPE.
  • Built a Python/PySpark-based marketing ROI optimization engine on AWS to allocate $500M+ in annual credit card marketing spend across acquisition channels, maximizing NPV under expense, account, and risk-weighted asset (RWA) constraints.
  • Developed a customer default prediction (PD) model using Logistic Regression, Decision Trees, Random Forest, CatBoost, and XGBoost; built vintage-level delinquency (DPD bucket) tracking to assess borrower risk behavior across loan products.
  • Engineered a hybrid Prophet-XGBoost vintage forecasting model (custom cash-flow engine) achieving under 3% MAPE for capital, interest, and reserve forecasting across products, territories, and branches; deployed via Tableau dashboards.
  • Built an NLP- and computer-vision-based OCR pipeline (Tesseract, EasyOCR, Hugging Face) to extract and validate data from 40,000+ scanned documents, detecting fraudulent claims and mismatched-item disputes at 98%+ accuracy; used Hadoop-based distributed processing to cut runtime from a day to 30 minutes.
  • Applied K-Means clustering on AWS/PySpark to segment millions of customer records for customer lifetime value (CLV) analysis.
  • Deliver model governance documentation (AMR, MDD, MCA) and mentor junior analysts on modeling and delivery standards.

Proof-of-Work

No completed sessions yet. Anything Manish finishes shows up here with the evidence behind it.

Proof of Education

B.Tech, Engineering

Indian Institute of Technology (IIT), Hyderabad · 2019 — 2023

A Proof CV — one profile, kept current, with the proof attached. Back to the portal