5 months ago
Hyderābād, IndiaSenior
Responsibilities
- Design precise classification and extraction prompts, including few-shot, chain-of-thought, and structured-output templates.
- Own prompt library versioning and rollback strategy.
- Create ground-truth datasets and automated evaluation pipelines measuring precision, recall, and field-level accuracy.
- Design multi-agent document-processing pipelines and integrate MCP servers and tool-use patterns.
- Build production-grade APIs and services around LLM capabilities using maintainable, testable architecture and well-structured Python.
- Contribute to CI/CD, observability, and deployment pipelines while owning architectural and technical-debt decisions.
- Act as the technical interface between business teams and the AI engineering layer, communicating constraints, system boundaries, and quality metrics.
Requirements
- Advanced prompt engineering and LLM evaluation framework experience.
- Experience designing ground-truth datasets for classification and extraction tasks.
- Experience with LLM APIs including OpenAI, Anthropic, and Azure AI.
- 8+ years of overall software engineering experience, preferably with Python and AWS.
- Experience with AI agent orchestration, Clean Architecture or DDD, REST API design, Git, CI/CD, and containerisation.
- Nice-to-have experience with document understanding or OCR tooling such as Textract.
- Nice-to-have experience with AWS services including EC2, SQS, S3, Lambda, and ECS.
- Nice-to-have experience with LangChain, LlamaIndex, NLP, or text classification.
Categories
About GHX
GHX (Global Healthcare Exchange) builds a cloud-based supply chain platform and B2B exchange that connects hospitals with medical suppliers for e-procurement, invoicing, content/contract and inventory management, and data synchronization. The company sells healthcare SaaS applications, network services, and analytics to providers, manufacturers, and distributors in North America and Europe. Founded in 2000 and headquartered in Denver, GHX is privately held and owned by Veritas Capital, and reports saving the industry over $5 billion since 2010.