Open to opportunities

Suraj Gupta

@surajgupta2

Message

Data Scientist blending financial risk modeling with practical AI and graph-based systems.

India

Message

What I'm looking for

I want to work as a Data Scientist applying ML to high-stakes problems—building models that improve accuracy and decision quality, with room to own projects end-to-end and collaborate across risk, data, and engineering teams.

I’m a Data Scientist focused on building reliable ML models for real-world decisioning—especially in financial risk modelling. At American Express, I used GMM clustering for PD stability analysis across 6 markets, redesigned the IFRS staging methodology to reduce false positives, and improved CCAR recovery and EAD modelling accuracy by 8.9% while reducing build time.

Across internships and projects, I’ve applied AI/ML to extraction, forecasting, and fraud use cases—building knowledge-graph retrieval with Neo4j and LangChain, generating Cypher queries with AutoGen for a Chainlit chatbot, and delivering end-to-end pipelines with XGBoost/RF and deep learning approaches like ANN. I also like challenging problems: I secured 2nd place out of 130 teams in an American Express challenge and ranked 1st in a collections forecasting hackathon by combining strong modelling with business-driven constraints.

Experience

Work history, roles, and key accomplishments

Current

Data Scientist

Current

American Express

Jul 2025 - Present (1 year)

Determined optimal origination tenure using GMM clustering-based PD stability analysis across 6 international markets. Redesigned IFRS SICR staging methodology to reduce false positives and developed CCAR recovery/EAD models improving accuracy by 8.9%.

Gaussian Mixture Model IFRS 9 IFRS SICR Staging CCAR EAD Recovery Models Model Development

Data Analytics Intern

EXL Services

Jul 2024 - Aug 2024 (1 month)

Enhanced fraud detection capabilities for UK banking data with a focus on accurate and efficient data handling using ML/DL. Addressed class imbalance with SMOTE, applied Robust Scaler and z-test feature selection, and built XGBoost and ANN models achieving F1=0.86 and AUC=0.92 with a 25% improvement in identification.

Fraud Detection Machine Learning Deep Learning SMOTE Robust Scaler XGBoost Model Evaluation

AI/ML Intern

Elucidata Data Consulting

Jan 2024 - Jun 2024 (5 months)

Extracted biomedical entities (diseases, drugs, genes) and relationships, storing results in a graph database using RAG techniques. Built a Neo4j + LangChain knowledge-graph retrieval system and developed AutoGen-based agents to generate Cypher queries integrated into a Chainlit chatbot.

RAG Biomedical Entity Extraction Knowledge Graphs Graph Databases Neo4j Lang Chain AutoGen Cypher Query Generation Chainlit Information Retrieval

Data Science Intern

Raymond Limited

Jul 2022 - Aug 2022 (1 month)

Designed an end-to-end machine learning pipeline to enhance cross-sell opportunities for trousers to shirt buyers. Engineered a 27-dimensional feature set on 1+ lakh records using XGBoost and Random Forest, improving F2-score from 0.72 to 0.84 and deployed the model to increase response rate by 3% on 10,000 customers.

Machine Learning Pipelines Feature Engineering XGBoost Random Forest Deployment

Education

Degrees, certifications, and relevant coursework

Indian Institute of Technology Kharagpur

BS/MS in Economics, Economics

2020 - 2025

Grade: CGPA: 8.43

Activities and societies: Minor in Mathematics & Computing; Micro specialization in AI & ML.

Pursued a BS + MS program in Economics at IIT Kharagpur from 2020 to 2025, achieving a CGPA of 8.43. Completed a minor in Mathematics & Computing and a micro specialization in AI & ML.