Generative AI Engineer

Hiring : AI Architect (7+ yrs) — Agentic & GenAI Systems

Location :

Bengaluru / Gurugram |

Full-time | AI Platforms & Architecture Team

About the Role

We’re looking for an

AI Architect

with a strong engineering background and proven expertise in

Agentic Frameworks

and

Generative AI platforms .

You’ll own the design and scale of production-grade

agentic systems , ensuring they’re secure, observable, and optimized for performance and cost on AWS.

This role demands a

hands-on architect

— someone who can debug code, model infrastructure, and lead developers through deep technical challenges.

⚙️

What You’ll Do

Architect and scale

Agentic AI systems

(planning, reasoning, and tool-calling) and

GenAI / RAG pipelines .

Design production architectures on

AWS (EKS, Bedrock, S3, SQS / SNS)

with a focus on cost, performance, and reliability.

Implement and optimize

multi-agent frameworks

such as

Semantic Kernel, LangGraph, AutoGen, or CrewAI .

Build retrieval and hybrid search pipelines with grounding, caching, and hallucination control.

Drive

observability

via OpenTelemetry + Datadog (metrics, tracing, SLIs / SLOs).

Set security, compliance, and data-governance standards for GenAI deployments.

Mentor engineers, conduct design reviews, and lead technical architecture discussions.

What We’re Looking For

7–10 years

in software / AI engineering with at least

4+ years in GenAI applications .

Agentic framework experience is mandatory

(Semantic Kernel, LangGraph, AutoGen, or CrewAI).

Expert in

Python (3.11+), FastAPI, asyncio, profiling, and pytest .

Proven experience in

architecting GenAI / RAG systems

using

AWS Bedrock , OpenSearch Serverless, or equivalent.

Deep understanding of

AWS cloud services , distributed systems, and observability.

Passion for building reliable, secure, and high-performing AI platforms at scale.

Bonus Points

Expertise in

multi-agent orchestration

and reasoning flows.

Hands-on experience with

vector databases

(Pinecone, Redis, pgvector).

Familiarity with

evaluation frameworks , rollout gates, and FinOps optimization.

Exposure to

policy-as-code (OPA) ,

feature management (LaunchDarkly) , or

NeMo Guardrails .