Website:
kasparro.com
Job details:
Full-time · Bengaluru · On-site
Experience: 2–3 years (including 1.5+ years shipping LLM/AI products to production)
Compensation: ₹10–15 LPA fixed + ESOPs of matching value (Total CTC ₹20–30 LPA)
About Kasparro
Kasparro builds the infrastructure that measures and improves how brands are represented inside AI engines such as ChatGPT, Claude, Gemini, Perplexity, and Google AI.
As AI engines become the layer buyers check before making decisions, the way brands are represented inside these systems has become increasingly important. We measure that representation, explain it, and help improve it.
Our platform is live, in production, and used by paying customers across India and the US.
Why this role exists
The measurement layer is live, and the next stage is more complex than what we've already built: continuous monitoring, systems that evaluate their own accuracy, and agents that carry memory across runs.
We're looking for an engineer who can take problems end-to-end, make sound technical decisions, and decide where model judgment ends and deterministic code begins. The surface you own is genuinely yours, with real responsibility from day one.
The role
You'll work on a production multi-agent system where LLM agents plan, extract, draft, and critique, built on a retrieval layer that reasons from evidence rather than memory.
Deterministic code computes every number and governs every stage. Knowing which part of a problem belongs to the model, and which belongs to code, is central to this role.
What makes it hard
- Deciding what belongs to model judgment and what must remain deterministic.
- Designing multi-agent systems that fail safely rather than silently.
- Solving production incidents by identifying root causes, not just symptoms.
What you'll own
- End-to-end product features, from design and data contracts to deployment, monitoring, and error recovery.
- Agent orchestration, including tool calling, structured output validation, multi-provider routing, and fail-closed handling.
- Retrieval systems, including semantic search, document ingestion, and retrieval quality.
- PostgreSQL schema design and production-safe migrations.
- Production health, including testing, evaluations, CI/CD, tracing, latency, cost monitoring, and incident response.
- Frontend development in React and Next.js for dashboards and report views, built backend-first and AI-assisted with Claude Code or Cursor.
What we're looking for
We're looking for engineers who have:
- Strong system design skills and the ability to explain technical trade-offs.
- 2–3 years of experience, including 1.5+ years shipping LLM/AI products to production.
- Experience with agent orchestration and retrieval systems.
- Strong Python, async programming, FastAPI, and PostgreSQL fundamentals.
- Experience measuring, testing, and operating AI systems in production.
- A backend-first mindset with the ability to build clean, working frontends using React/Next.js and AI assistance.
- Engineering depth, shipped work, and the ability to learn the domain quickly matter more.
Your first 90 days
You'll begin by learning the domain and the codebase while shipping reviewed production work.
As your understanding grows, you'll take ownership of complete features before progressing to owning an entire module, including its architecture, failure modes, testing, and production reliability.
You'll work directly with the founder and the engineering team as the platform continues to grow.
Why join us
- Work on a production multi-agent system with paying customers, not a prototype.
- Own critical backend and agent systems from day one.
- Work directly with repeat founders on a team that ships every week, solving engineering problems at the intersection of AI and software systems.
How to apply
Apply through LinkedIn Easy Apply.
You can also email your application to hazel@kasparro.com, CC grandmaster@kasparro.com.
Subject: "Full Stack Engineer II — [Your Name]".
Include: Your CV or LinkedIn profile
Click on Apply to know more.