SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
SS&C is a leading provider of AI-powered technology and services for financial services and healthcare organizations. With 27,000+ employees across 35 countries and 20,000+ client organizations relying on their expertise, SS&C is headquartered in Windsor, Connecticut.
In this Advanced Prompt Engineer role based in Hyderabad (hybrid), you will be responsible for designing, building, and optimizing sophisticated prompt structures that power core business applications. Your work will span the full lifecycle of prompt engineering: constructing Few-Shot, Chain-of-Thought, and ReAct prompts; integrating guardrail frameworks (NeMo Guardrails and custom logic) to enforce safety constraints and prevent improper outputs; and systematically evaluating prompt performance against business objectives.
Key responsibilities include:
- Construct, maintain, and systematically evaluate complex prompt structures for core business tasks
- Integrate program-level guardrail frameworks to enforce safety rules and prevent improper outputs
- Optimize system instructions and prompt context formatting to minimize token consumption while maintaining precision
- Implement programmatic prompt tuning techniques (using DSPy or custom evaluation scripts) to refine system outputs over baseline benchmarks
- Ensure outputs consistently adhere to strict target schemas (JSON, XML, Pydantic objects) for downstream API ingestion
- Partner with Business Analysts to convert operational guidelines into precise model system prompts
You will work with foundation models including GPT-4, Llama, Claude, and Mistral, optimizing for both performance and cost efficiency. This is a technical role requiring strong software engineering fundamentals and deep understanding of how large language models behave in production environments.
REQUIREMENTS:
- 3–5 years of experience in software engineering, computational linguistics, or technical prompt development
- Deep knowledge of foundation model behaviors (GPT-4, Llama, Claude, Mistral), context window dynamics, and tokenization mechanics
- Strong Python coding skills to programmatically generate, test, and evaluate prompt performance at scale
- Hands-on experience with output validation libraries (Pydantic, Instructor) and guardrail systems
- Strong technical discipline around edge-case mapping, deterministic output formatting, and cost optimization
