Senior Software Engineer - Agentic AI

Yubi; Bengaluru; On-site; INR 1,800,000 - 3,000,000; Full time

Job summary

Yubi is looking for a Senior Software Engineer on the Agentic AI team to build and deploy intelligent agents that automate debt-market workflows. You will own features end-to-end—writing Java backend services and wiring LLM-powered reasoning chains that turn financial processes into auditable pipelines.

You will mentor peers, contribute to architecture decisions, and collaborate across teams while ensuring production robustness with tests, observability, and CI/CD on AWS.

Qualifications

• 3–5 years of full-stack production experience.
• Strong Java fundamentals, Spring Boot APIs.
• Experience with LLMs and vector databases is desired.
• Familiarity with AWS and Docker environments.

Responsibilities

• Build multi-step AI agent pipelines using LangChain or similar.
• Connect LLMs to internal APIs, GST, and MCA sources.
• Design and maintain RAG pipelines with vector databases.
• Write deterministic tests and participate in code reviews.
• Instrument observability and deploy services on AWS with CI/CD.

Skills

Core Java

Spring Boot

LLM API integration

LangChain/LangGraph

Vector databases

PostgreSQL

Redis

AWS

Docker

CI/CD (GitHub Actions)

Multithreading

Tools

Docker

GitHub Actions

LangChain

Pinecone/Weaviate/pgvector

Job description

Yubi (formerly CredAvenue) is redefining global debt markets by freeing the flow of finance between borrowers, lenders, and investors — the world's possibility platform for the discovery, investment, fulfilment, and collection of any debt solution. In March 2022, we became India's fastest fintech unicorn with a $137M Series B. Our platforms — Yubi Credit Marketplace, Yubi Invest, Financial Services Platform, Spocto, and Corpository — serve 17,000+ enterprises and 6,200+ investors, facilitating over ₹1,40,000 crore in debt volumes. Backed by Insight Partners, Sequoia Capital, Dragoneer, B Capital, LightSpeed, and Lightrock.

ROLE OVERVIEW

As a Senior Software Engineer on Yubi's Agentic AI team, you will build and deploy intelligent agents that automate complex debt-market workflows — from credit underwriting to collections orchestration. You will own features end-to-end: writing Java backend services, and wiring LLM-powered reasoning chains that turn financial processes into reliable, auditable pipelines. You will contribute to architecture decisions, mentor peers, and help raise the bar for engineering quality across the team.

KEY RESPONSIBILITIES

Agentic AI Development

• Build multi-step AI agent pipelines using LangChain, LangGraph, or custom orchestration to automate lending and compliance workflows.
• Implement tool-use and function-calling layers connecting LLMs to internal APIs, bureau data, GST, and MCA sources.
• Design and maintain RAG pipelines with vector databases (Pinecone, pgvector, Weaviate) for document intelligence and semantic search.
• Write reliable prompt pipelines with structured output parsing, retry logic, and guardrails for production robustness.
• Build human-in-the-loop checkpoints — approval gates, anomaly escalation, and audit trails — for high-stakes decisions.

Quality & Collaboration

• Write deterministic tests for non-deterministic agent behaviour using golden datasets and regression suites.
• Participate actively in code reviews, architectural discussions, and Agile ceremonies.
• Instrument agents with observability tooling (LangSmith, Datadog) to monitor token usage, latency, and failure modes.
• Containerise and deploy services on AWS using Docker and CI/CD via GitHub Actions.

Requirements

REQUIRED SKILLS & QUALIFICATIONS

Technical

• 3–5 years of full-stack engineering experience in production environments.
• Solid Core Java fundamentals — OOP principles, collections framework, multithreading, and exception handling.
• Proficiency in Spring Boot for building scalable REST APIs and microservices.
• Hands-on experience integrating LLM APIs: OpenAI, Anthropic, Gemini, or open-source equivalents.
• Working knowledge of agent frameworks — LangChain, LangGraph, AutoGen, or similar.
• Familiarity with vector databases and RAG pipeline design.
• Experience with PostgreSQL, Redis, and cloud-native AWS services (Lambda, S3, ECS).
• Comfortable with Docker and CI/CD pipelines.