Agentic Software Engineer-II

Equifax, Inc.; Thiruvananthapuram; Presencial; INR 2.000.000 - 4.000.000; Jornada completa

Descripción de la vacante

Equifax, Inc. is seeking a hands-on Agentic AI Engineer to join our enterprise AI Platform team in Trivandrum. You will build, test, and deploy autonomous AI agents capable of reasoning, multi-step planning, tool utilization, and execution in enterprise environments.

In this role, you will optimize prompts, implement stateful workflows, integrate APIs, and ensure production-grade reliability with clean, performant code and robust testing, tracing, and observability.

Formación

• 3-6 years of software engineering with 1+ years building production-grade LLM apps.
• Proficiency in Python and LLM orchestration frameworks.
• Experience writing JSON schemas for function calling and API wrappers.
• Experience with observability, tracing, and performance monitoring.

Responsabilidades

• Implement stateful, multi-step agent behaviors and optimize prompts.
• Develop robust API wrappers and JSON schemas for tool-calling and integration.
• Build context retention systems and memory buffers for multi-turn interactions.
• Create high-throughput ingestion pipelines and improve search relevance.
• Implement error handling, retries, and self-correction mechanisms.
• Write deterministic tests and use tracing tools to debug reasoning paths.

Conocimientos

Python

LLM orchestration frameworks

API design

JSON schema design

Testing and QA

CI/CD

Observability and tracing

Herramientas

LangGraph

LangChain

AutoGen

LlamaIndex

LangSmith

Phoenix

Docker

Descripción del empleo

Equifax is looking for a hands-on Agentic AI Engineer to join our enterprise AI Platform team in Trivandrum. Working under the guidance of Lead Agentic Architects, you will build, test, and deploy autonomous AI agents capable of reasoning, multi-step planning, tool utilization, and execution in enterprise environments. In this role, you will transition AI prototypes into resilient production systems focusing on writing clean, performant code, optimizing LLM prompts, integrating APIs, and managing stateful agentic workflows.

What You’ll Do

• Implement stateful, multi-step agent behaviors using frameworks like LangGraph, AutoGen, or LlamaIndex. Optimize prompts using techniques like Chain-of-Thought (CoT) and ReAct.
• Develop resilient API wrappers and strict JSON schemas for LLM tool-calling/function-calling, connecting agents safely to internal databases, CRMs, and enterprise systems.
• Build context retention systems using short-term memory buffers and long-term vector/relational storage for multi-turn user interactions.
• Build and maintain high-throughput ingestion pipelines; optimize search relevance using hybrid search, embeddings tuning, and re-ranking algorithms.
• Implement error handling and retry mechanisms for rate limits, API failures, and hallucinations, enabling agents to self-correct or cleanly elevate to human operators.
• Write deterministic software tests and LLM-as-a-judge evaluations. Utilize tracing tools (e.g., LangSmith, Phoenix) to debug reasoning paths, eliminate infinite loops, and optimize latency.

What Experience You Need

• 3-6 years of core software engineering experience, with at least 1+ years hands-on experience building and deploying production-grade LLM applications, RAG pipelines, or Agentic workflows.
• Strong proficiency in Python (FastAPI, Pydantic, Asyncio) and hands-on experience with LLM orchestration frameworks (LangGraph, LangChain, AutoGen, or LlamaIndex).
• Experience writing JSON schemas for function calling, building REST/gRPC API wrappers, and working with vector databases
• Experience using observability platforms for tracing agent execution, latency monitoring, and prompt evaluation.
• Solid background in CI/CD, Git workflows, unit/integration testing, Docker containerization, and writing clean, scalable software.

What Could Set You Apart

• Hands-on experience with cloud-native AI services on GCP (Vertex AI, ADK, Model Armour)
• Exposure to semantic guardrails framework (e.g., NeMo Guardrails, Guardrails AI).
• Familiarity with fine-tuning techniques, embedding customization, or open-source local model deployment