Product Specialist (T)

Forward Deployed Engineer

Archetype: Builder · Learner · Pair programmer

Role Summary

You are an on-the-ground builder who writes real production code inside client environments from day one. You come in hungry, move fast, and turn ambiguous problems into working prototypes within hours — not weeks. Your credibility comes entirely from working software, not slides.

What You Will Do

•     Embed directly at client sites to prototype and deploy agentic AI workflows using frameworks such as LangGraph, CrewAI, AutoGen, or AWS Bedrock Agents — shipping working code, not slide decks.

•     Build RAG pipelines end-to-end: chunking strategies, vector store configuration (Pinecone, pgvector, Weaviate), retrieval tuning, and response evaluation.

•     Instrument LLM-powered applications with observability tooling (LangSmith, Braintrust, Arize) so clients can see exactly what their agents are doing in production.

•     Participate actively in daily client stand-ups and technical reviews, communicating clearly about progress, blockers, and trade-offs with both engineers and business stakeholders.

•     Rapidly iterate on prototypes based on user feedback — from zero to demo in 24–48 hours is the expectation, not the exception.

•     Document deployment architectures, prompt engineering decisions, and integration patterns so knowledge persists after you rotate off an engagement.

•     Contribute reusable agent templates and accelerators to Cognizant's internal AI toolkit between engagements.

Technical Foundation

•     Strong Python; basic TypeScript / JavaScript

•     REST API design and integration

•     Git, CI/CD basics, containerisation (Docker)

•     SQL and at least one cloud platform (AWS / Azure / GCP)

•     Hands-on LLM experience (OpenAI, Anthropic, Gemini APIs)

GenAI / Agentic AI Requirements

•     Has built at least one end-to-end RAG or agent application — personal projects count as strongly as work experience

•     Understands prompt engineering, few-shot design, and chain-of-thought prompting

•     Familiar with agentic orchestration concepts: tool use, memory, planning loops

•     Knows how to evaluate LLM output quality — even informal logging or manual review frameworks

What Makes You Stand Out

•     You have shipped something real with AI — a GitHub repo, a side project, a hackathon win — not just certificates

•     You are comfortable being wrong in front of a client and pivoting immediately

You ask 'what does done look like?' before writing a single line of code