Applied AI Engineer

Strategic Talent Partner; India; On-site; INR 2,500,000 - 3,500,000; Full time

Job summary

An AI Solutions Company based in India is seeking an experienced developer skilled in Python or JavaScript to design and deploy advanced multi-agent pipelines for credit analysis. The ideal candidate will have over 10 years of experience in production-grade AI systems, LLM Ops, and data engineering, as well as a strong understanding of prompt engineering and multi-agent control. This position offers the opportunity to work at the forefront of AI technology.

Qualifications

• 10+ years in Python or JS development.
• Proven experience with production-grade Agentic AI or reasoning systems.
• Strong background in LLM Ops, data engineering, and cloud platforms (AWS/Azure).

Responsibilities

• Design and deploy advanced multi-agent pipelines for credit analysis.
• Optimize inference and prompt chains using frameworks like DSPy, GEPA, and LangChain.
• Implement reasoning techniques and evaluation systems.

Skills

Python development

JavaScript development

Production-grade AI systems

LLM Ops

Data engineering

Cloud platforms (AWS/Azure)

Prompt engineering

Multi-agent control

Job description

Responsibilities

• Design and deploy advanced multi‑agent pipelines for credit analysis.
• Optimize inference and prompt chains using frameworks like DSPy, GEPA, and LangChain.
• Implement reasoning techniques (CoT, ToT, GoT) and evaluation systems (LLM‑as‑a‑judge).
• Develop ETL workflows, data pipelines, and observability dashboards.
• Manage memory, retrieval, and orchestration in multi‑agent systems.
• Maintain cloud‑based AI services and integrate with DevOps pipelines.

Qualifications

• 10+ years in Python or JS development.
• Proven experience with production‑grade Agentic AI or reasoning systems.
• Strong background in LLM Ops, data engineering, and cloud platforms (AWS/Azure).
• Expertise in prompt engineering, RAG, HITL, and multi‑agent control.

Nice‑to‑have

• Experience with RLHF/RLAIF or fine‑tuning (LoRA/QLoRA).