Role Overview
This is the pod’s most critical hire. You set the technical direction for every piece of AI work at Kissht - architecture, integration patterns, evaluation infrastructure, and the call between vendor APIs and self-hosted models. When someone asks “how do we do AI engineering here?”, the answer is yours to give.

Responsibilities - building and delivery
AI platform architecture: LLM integrations, RAG pipelines, agent frameworks, context engineering, evaluation harnesses, observability.
Build-vs-buy technical calls, in partnership with other stakeholders.
Engineering standards for AI work across Kissht: how we prompt, evaluate, monitor, and iterate.
Technical vendor evaluations: API testing, latency and accuracy benchmarks, integration complexity, red-flag surfacing.
Mentoring the two SDE-2 Builders and the Data Engineer. You set the quality bar.

What success looks like - first 90 days
The pod’s AI architecture is documented, reviewed, and being followed in active builds.
The evaluation harness used by the pod is operational and used in at least one live vendor eval.




The two SDEs are productive and unblocked technically.

What success looks like - first 180 days
One in-house AI orchestration layer is in production, owned by the pod, and integrated with at least two Kissht systems.
The build-vs-buy decision flow has been used on at least three vendor decisions, with documented rationale.
AI engineering standards are visible to non-pod engineers and at least one team outside the pod has adopted them.

Must-haves
7–8 years of backend engineering experience. At least 2 years building LLM-based products in production.
Deep familiarity with at least one production-grade LLM ecosystem and its orchestration patterns.
Hands-on experience with RAG architecture, vector databases, evaluation pipelines, and context engineering.
Solid systems thinking: prompt → data pipeline → model → integration → monitoring, end to end.
Strong Python. Comfortable reading and writi

📌 Tech Lead (Mumbai)
🏢 Kissht
📍 Mumbai