AI Engineer Copy 01

Accenture PLC; Petaling Jaya; On-site; MYR 120,000 - 180,000; Full time

Job summary

Accenture plc is seeking a software engineering talent to join the team, focusing on production-grade Python, LLM-based AI systems, retrieval engineering, and trustworthy AI within healthcare/pharma tech solutions.

The role emphasizes multi-agent orchestration, RAG, graph-grounded outputs, and compliance with GxP/CSV data governance while delivering on regulated delivery constraints.

Qualifications

• Production-grade Python development experience.
• Experience with LLMs, multi-agent orchestration and RAG.
• Experience in retrieval engineering, embeddings and document parsing.
• Knowledge of trustworthy AI evaluation and human-in-the-loop design.
• Cloud & MLOps on AWS with observability tooling.
• Familiarity with GxP/CSV data governance constraints.
• Experience in spec-driven development.

Skills

Python

LLM AI

Retrieval Eng

Trustworthy AI

AWS MLOps

GxP/CSV

Spec-driven Dev

Tools

LangGraph

LangChain

RAG

OpenSearch

Neptune

SageMaker

LangSmith

CloudWatch

Job description

Client & Project: We are seeking a new talent to join the Software Engineering team where you will have the opportunity to collaborate in the project .The client is a company specializing in healthcare and pharmaceuticals with a focus on innovative medical solutions.

Skills required:

• Python engineering. Production-grade Python
• Applied LLM & agentic AI. Multiagent orchestration (LangGraph / LangChain), RAG and graph-grounded RAG, structured/typed outputs, and tool use via MCP.
• Retrieval engineering. Hybrid vector-plus-graph retrieval, embeddings, layout-aware document parsing, and entity resolution.
• Trustworthy AI & evaluation. LLM-as-judge plus human-review sampling, groundedness and citation faithfulness, hallucination control, and human-in-the-loop design.
• Cloud & MLOps on AWS. AgentCore, OpenSearch, Neptune, and SageMaker, with observability via LangSmith and CloudWatch/OpenTelemetry.
• Regulated delivery. Comfort building under GxP/CSV and formal data-governance constraints
• Spec-driven development