Data Science Intern

Data Science Intern → Full-time (Core Models)
Affluense.ai | Bengaluru (BHIVE, in-office) | 3-month internship with a path to full-time
Duration: 3 months, full-time internship
Conversion: Full-time Data Scientist offer based on performance at the end of month 3
Start date: Immediate
ABOUT AFFLUENSE.AI
Affluense.ai is building India's wealth intelligence layer. Our AI-powered B2B platform helps wealth managers, private bankers and family offices find, understand and engage High Net Worth Individuals (HNIs).
We're live with some of India's largest private wealth and wealth management firms. We're backed by Zeropearl VC, alongside angel investors Kunal Shah (CRED) and Pravin Jadhav (Dhan).
WHY THIS ROLE EXISTS
The models behind Affluense are the product: wealth scoring, matching people across sources, segmentation and prospect ranking. We want an early-career data scientist who will grow into owning a piece of them.
This is a classical ML role: supervised and unsupervised models on messy, real-world structured data. It is not an LLM, prompt-engineering or RAG role.
You'll work directly with our CEO, who led data science at HealthifyMe, Mobile Premier League, BrowserStack and Simpl (where he built alternate-data credit models). You'll also work with our CTO, who has 18+ years in engineering and was Head of Engineering at Edelweiss/Nuvama Wealth.
WHAT YOU'LL WORK ON
- Wealth signal modeling: build features and supervised models (regression, classification, gradient boosting) that estimate wealth from indirect signals like directorships, property records and company filings
- Segmentation and clustering: use unsupervised methods (k-means, hierarchical clustering, PCA) to group HNIs into segments customers can act on
- Entity matching: match and deduplicate the same person across sources with inconsistent names, dates and identifiers
- Data mining: dig into large, messy datasets, find what actually predicts wealth, and validate it before it reaches a model
- Model evaluation: design validation when there's no clean ground truth, and explain results in plain language
- Customer-facing analysis: turn model outputs into clear Excel analyses for the business and sales teams
MUST-HAVES
- At least one prior data science internship or professional role where you built traditional ML models, either supervised (regression, classification, tree-based) or unsupervised (clustering, dimensionality reduction)
- Solid basics: statistics, probability, feature engineering, bias–variance, cross-validation, and picking the right metric (precision/recall, AUC, RMSE)
- Hands-on Python: pandas, NumPy and scikit-learn
- Strong Excel: pivot tables, lookups (VLOOKUP/XLOOKUP/INDEX-MATCH), conditional logic, clean summaries
- Curiosity about the data: you ask why a pattern exists before you model it
- Clear communication: you can explain a model and its limits to a non-technical person
GOOD TO HAVE
- SQL
- XGBoost / LightGBM
- Exposure to fintech, credit risk, fraud, BFSI or wealth data
- Work on record linkage, deduplication or network analysis
- A degree in engineering, statistics, mathematics or economics, or equivalent proof of skill
WHY JOIN
- Your models go live with India's leading wealth management firms
- Direct mentorship from founders with deep data science and BFSI backgrounds
- A clear path to full-time in an early, funded team
- Hard, unusual problems: estimating wealth without clean ground truth

Work arrangement
No