Role : AI Engineer
Experience : 1–3 years
Location :
Mumbai / Bengaluru / Gurgaon (Hybrid : 3 days / week in office)
Remote option for exceptional candidates.
About the Role
We’re building production-grade AI workflows and agentic applications that power real user experiences. As an AI Engineer, you’ll ship features end-to-end—from prompt design and evaluation to scalable backend integration—working closely with product and engineering.
What You’ll Do
- Design, build, and iterate LLM-powered workflows (retrieval, routing, tool use, function calling, multi-step agents).
- Implement agentic apps that plan, call tools / APIs, and maintain state across tasks.
- Build RAG pipelines : data ingestion, chunking, embeddings, indexing, and latency-optimized retrieval.
- Own prompt engineering & evaluation (A / B tests, guardrails, metrics like latency, cost, quality, safety).
- Productionize models / workflows with observability (traces, tokens, failures), cost controls, and fallbacks.
- Ship backend services and APIs (e.g., Python / FastAPI) integrating with data stores and vector DBs.
- Collaborate with PM / Design to translate requirements into reliable, user-facing features.
Must-Have Skills
Hands-on with LLMs (OpenAI, Claude, Llama, etc.) and orchestration frameworks (LangChain, LlamaIndex, or custom).Python proficiency; building RESTful services, writing clean, tested code.Experience with RAG, vector databases (Pinecone, Weaviate, FAISS, Qdrant), embeddings.Understanding of agent patterns (tool calling, planning / execution, memory) and workflow engines.Familiarity with prompt design, safety / guardrails, and evaluation frameworks.Basics of cloud & deployment (AWS / GCP / Azure), Docker, Git, CI / CD.Strong debugging mindset and bias to ship.Nice-to-Have
FastAPI / Flask, Celery / queues; streaming UIs.Model fine-tuning / LoRA, dataset curation, prompt-cache strategies.Monitoring / tracing (LangSmith, Weights & Biases, OpenTelemetry).Frontend basics (React / Next.js) to collaborate on UX.Data privacy, PII redaction, and security best practices.Success MetricsReduction in latency / cost per request; improvement in answer quality scores.Workflow reliability (timeouts, retries, fallbacks) and on-call readiness.Speed of iteration from spec → prototype → production.