Senior Data Scientist

About the Role

NIRA is a consumer credit company building a financial brand for Middle India. Our first product is credit, receiving 5000 new loan applications daily from across 4000 cities in India, and growing 20%+ MoM. We offer credit limits via our app, with money reaching bank accounts within 24 hours. Since our launch in 2018, we have achieved product market fit and are now entering our next phase of growth. We are looking for pragmatic data scientists to utilise their data science and engineering expertise to build decision engines, enabling access to credit for millions of Indians. You will work closely with the founders, underwriting & collections team to achieve these objectives.

Roles & Responsibilities

• Own end-to-end data science projects from the exploration phase to implementation in production to post-production iterations.
• Improve our infrastructure for training and deploying machine learning models, both batch and real-time.
• Consolidate datasets comprising a range of structured and unstructured data; get involved in ETL process if required.
• Develop contextual domain knowledge around the problem by interacting with a range of stakeholders.
• Build explanatory as well as predictive models for fraud detection & credit risk assessment.

Skills & Qualifications

• Strong proficiencies in various statistical and non-statistical approaches to machine learning.
• 3+ years experience dealing with large real-world datasets, running an iterative ML pipeline in a production environment.
• Substantial experience building mission-critical systems.
• You should be technical and be able to quickly pick up new engineering skills needed.
• You have a strong framework to differentiate a good model from the bad.
• You are curious to uncover the story behind the data.
• Domain experience in credit and risk is highly preferred.

To be successful in this role, you would need to be highly analytical, have the ability to synthesise information well from disparate sources, and be extremely curious about the problem! In return, you would get to work on proprietary, hard to reproduce, rich financial and behavioral datasets and ship models that will be a force multiplier for NIRA and its customers.

Our Tech Stack

The primary interface for our 25000 daily active users is our Android app written in Java. On the backend we have 100+ microservices with serverless deployment that handle the entire user services and workflow management from onboarding, data collection, decision engine, document processing, disbursals and collections. This is complemented with an internal Node.js application for Operations workflow management. Our data science modules & decision engines are written in Python and deployed over serverless architecture as well. We have DynamoDB as our primary database, and AWS Glue based ETL architecture for our secondary database. Querying on secondary database is via Athena.