Principal Engineer

Krazy Bee Services; Bengaluru; On-site; INR 6,000,000 - 9,000,000; Full time

Job summary

KreditBee is seeking a Principal Engineer, Architecture & AI-Native Engineering in Bengaluru. You will lead a hands-on team of about 35 engineers, shaping the architecture of our lending platform and the AI-powered delivery path.

You will write production code on critical paths, establish design standards for reliability and observability, and mentor engineers across squads while collaborating with the SVP Engineering. On-site role.

Qualifications

• 12+ years building software at scale.
• Experience as hands-on architect of high-scale distributed systems.
• Contributed to AI-enabled engineering and tooling.

Responsibilities

• Own target architecture for KreditBee's app stack and data contracts.
• Lead a 3–5 engineer team building internal tooling and golden paths.
• Write production code on critical paths and drive design docs.
• Define AI-native delivery practices and ensure auditable AI usage.

Skills

Hands-on architecture

Team leadership

AI-native engineering

Distributed systems

Hands-on coding

Tools

MCP servers

CI integrations

Agent tooling

Job description

Principal Engineer, Architecture & AI-Native Engineering

Engineering • Office of the SVP Engineering • KreditBee

12+ years | Hands-on architect, leads a team of 35 | Full-time • On-site, all days • Bengaluru

About the role

KreditBee runs a regulated consumer-lending platform for millions of customers, spanning origination, servicing, collections, payments, and KYC. We are rebuilding how our ~150 engineers design, ship, and run software with AI, and we need a hands-on architect to lead that change.

This role owns two things that reinforce each other: the application architecture across our lending stack, and the engineering practices that let AI do most of the building. Agents move fastest in systems with clean boundaries, explicit contracts, strong tests, and good observability. You will shape our architecture to be AI-buildable and set the practices that put it to work.

You report to the SVP Engineering, lead a small team that builds AI-native tooling and golden paths, and work across every engineering team. The person we want has spent years as a hands-on architect of high-scale systems and has, in the last couple of years, rebuilt their own way of working around AI. You ship in days what used to take a team a sprint, tests and docs included, and you can show us how.

What youll own

Application architecture

• Own the target architecture for KreditBees application stack: service boundaries, data ownership, integration contracts, and a sequenced migration path from where we are today.
• Set standards for reliability, scale, security, and observability; run design review for significant changes and make the hard calls on build-vs-buy and tech debt.
• Write production code on critical paths and author the design docs that matter. You lead from the keyboard.

AI-native delivery

• Build the path from PRD to production with agents doing most of the work spec, code, tests, review, and deploy — while engineers own intent, judgement, and approval.
• Create golden paths and reference implementations that teams adopt because they are faster; measure adoption and iterate.
• Guard against the known failure modes of agentic coding: context loss, tests written to pass, review overload, and insecure dependencies.

Agentic operations

• Move on-call from manual runbooks to agent-assisted operations, where agents triage incidents, correlate logs, traces, and recent deploys, and propose remediation.
• Define what agents may read and change in production, where human approval is required, and how every action is logged.

Safe by design, for a regulated lender

• Keep customer data out of prompts, context windows, and agent logs; decide which models and providers may see our source code and data.
• Make every AI-generated change auditable: who prompted it, who reviewed it, and who approved it.

Your team

You will hire and lead 3–5 engineers building internal tooling: agent workflows, CI integrations, MCP servers, and golden-path templates. You will partner closely with engineering managers and tech leads across the org, and with our AI Platform team, whose models and gateway your tooling builds on.

Who you are

• 12+ years building software, with several as the hands-on architect of high-scale distributed systems in consumer tech, fintech, payments, travel, e-commerce, or a large GCC.
• You still write code every week, and you hold a high bar for design, testing, and operability.
• AI has changed how you personally work. You use agentic coding tools daily, you have built tooling for yourself or your team — MCP servers, CI-integrated agents, review bots, evals for your own workflows — and you can explain what works, what breaks, and why.
• You have run production systems with on-call and have strong views on which parts of operations agents should handle.
• You lead through example and persuasion. You have driven a technical change across many teams, and you have hired and grown engineers.
• Nice to have: lending, payments, or other regulated domains; platform engineering or developer-productivity work; open-source contributions or public writing on AI-assisted engineering.

What success looks like in year one

Delivery speed

PR-to-production lead time and PR cycle time drop sharply across the org; a typical feature moves from approved PRD to production in days.

Golden paths

Most teams build new services and features on the paths your team provides.

Operations

Agents triage a meaningful share of on-call incidents first, and MTTR trends down.

Architecture

A documented target architecture with a sequenced migration plan, and the first major moves shipped.

Governance

AI use across the SDLC is auditable and meets our data-protection and regulatory obligations.

How we’ll assess

Our loop includes a live working session: we give you a small, realistic PRD and a few hours with your own AI setup, and you ship it with tests. We also go deep on an architecture you designed and ran in production — what you built yourself, what broke, and what you would change. We care most about seeing how you work.

Why this matters.

This role sets how KreditBee builds and runs software for years to come.