AI Scientist - Credit Risk & Agentic AI

Job Description:

We are looking for an AI Scientist Credit Risk & Agentic AI to join our client in the AI and decision intelligence space.

This role is ideal for someone who combines credit-risk analytics, applied machine learning, LLM reasoning, evaluation design, and business judgment. This is not a generic prompt-engineering role. The selected candidate will help design, test, evaluate, and improve AI agents that analyze credit and underwriting data in a controlled, evidence-cited, human-in-the-loop environment.

Key responsibilities:

- Design AI reasoning frameworks for credit-risk and underwriting use cases.

- Translate underwriting judgment into structured, testable, and auditable AI workflows.

- Work on bureau/tradeline analysis, borrower-risk signals, shadow debt patterns, data-integrity flags, and score reconciliation.

- Design prompts, reasoning flows, evaluation criteria, and structured outputs for LLM-based underwriting agents.

- Define evaluation frameworks for underwriter alignment, incremental signal, hallucination risk, data-integrity detection, and decision quality.

- Compare AI outputs against historical underwriter decisions, credit scores, model outputs, and repayment/loss outcomes.

- Partner with AI engineers, data engineers, credit-risk experts, and platform teams to build governed AI underwriting solutions.

- Participate in internal and client-facing discussions with senior stakeholders.

Requirements:

- 5+ years of experience in data science, credit-risk analytics, underwriting analytics, risk modeling, or applied machine learning.

- Strong understanding of credit-risk concepts such as bureau data, credit scores, scorecards, delinquency, utilization, inquiries, tradelines, repayment behavior, loss/default outcomes, and underwriting policy.

- Strong Python and SQL skills.

- Experience with statistical modeling, machine learning, model validation, or analytical backtesting.

- Familiarity with LLMs, prompt design, structured outputs, RAG, or agentic AI systems.

- Ability to define evaluation frameworks and interpret model/AI outputs critically.

- Strong communication skills and ability to work with business, credit, data science, and engineering teams.

- Willingness to work from the Bhubaneswar office.

Preferred:

- Experience in fintech, lending, NBFC, banking, credit bureau, SME lending, consumer lending, merchant cash advance, or commercial underwriting.

- Experience with consumer bureau or commercial bureau data.

- Exposure to Claude, OpenAI, AWS Bedrock, LangChain, LangGraph, LlamaIndex, or similar tools.

- Familiarity with Pydantic, JSON schemas, evaluation harnesses, model governance, explainability, or human-in-the-loop AI workflows.

- Prior consulting or client-facing experience.

Benefits:

- Competitive compensation aligned with market standards, based on experience and capability.

- Opportunity to work on cutting-edge AI applications in real financial-services decisioning.

- Exposure to a high-impact North American fintech AI underwriting programme.

- Work on practical, governed AI systems rather than generic chatbot demos.

- Collaborate with experienced data science, risk, AI, and engineering professionals.

- Opportunity to build expertise in agentic AI, LLM evaluation, credit-risk analytics, and human-in-the-loop decisioning.

- Candidates from other cities are welcome to apply if they are open to relocating to Bhubaneswar.