Streamline Healthcare Solutions is a high growth technology company that delivers web-based software for healthcare organizations. The Lead AI Software Engineer role focuses on designing and delivering AI/ML solutions for the healthcare industry, including LLM-powered applications and predictive models, while ensuring compliance with HIPAA regulations.
Responsibilities:
- Co-design AI solutions with the AI Architect and own product-level solutioning and delivery within enterprise AI architecture, standards, and governance
- Lead end-to-end implementation for your product/squad: RAG over EHR/claims/clinical text using embeddings and vector search
- Model development and training/fine-tuning (LLMs and classical ML), evaluation frameworks, and guardrails (hallucination reduction, safety, PII/PHI handling)
- Production deployment with containerization/orchestration, and GPU-aware inference where applicable
- Establish and operate MLOps: experiment tracking and model registry, CI/CD for ML, canary/A/B testing, and monitoring for latency, accuracy, drift, bias, and cost
- Own reliability, security, and cost for your product’s AI services: define SLOs, participate in on-call/incident response, manage token/GPU budgets, and optimize prompts, embeddings, caching, and indexing
- Build and maintain data pipelines (e.g., Spark/Databricks or equivalent) and ensure robust SQL Server performance and data quality; collaborate with DBAs and data engineers using SSMS
- Ensure HIPAA compliance and Responsible AI practices across development and operations; partner with security and compliance to meet policy requirements
- Collaborate with product management, domain experts, and compliance to translate requirements into safe, reliable, high-impact AI services
- Conduct reviews emphasizing code quality, experiment rigor, reproducibility, and evaluation discipline; mentor engineers and data scientists
- Participate in architecture reviews; propose improvements and contribute reusable components (RAG templates, evaluation harnesses) back to the shared AI platform
- Leverage Microsoft Copilot and GitHub Copilot to improve developer productivity, code quality, and documentation, aligning with organizational governance
Requirements:
- Bachelor's degree in Computer Science, Information Technology, Computer Information Systems, Health Informatics, or a related field
- Healthcare domain experience: either clinical (e.g., care delivery, clinical documentation, quality) or revenue cycle management (e.g., coding, claims, denials, prior auth)
- 10+ years in software engineering; 5+ years building and shipping ML/AI solutions; 2+ years leading AI/ML initiatives or teams
- Azure AI Foundry (Azure AI Studio) knowledge for developing, evaluating, and operationalizing LLM solutions; familiarity with Azure AI resources and deployment patterns
- Proficiency with SQL Server Management Studio (SSMS) for SQL development, performance tuning, and troubleshooting; strong T-SQL fundamentals
- Proficiency with Visual Studio and experience integrating AI services into .NET/C# applications or services where needed
- Hands-on experience using OpenAI or Anthropic models (e.g., GPT-4.x/4o, Claude 3.x), including prompt engineering, function/tool calling, and evaluation
- Experience with Microsoft Copilot and GitHub Copilot in professional workflows (coding assistance, test generation, documentation) with awareness of usage policies and data boundaries
- Strong Python and ML ecosystem: PyTorch/TensorFlow, transformers/Hugging Face, embeddings, LLM orchestration (e.g., LangChain or LlamaIndex), and vector databases (e.g., FAISS, Azure AI Search, Pinecone)
- MLOps: MLflow/W&B (or equivalent), Docker, Kubernetes, CI/CD for ML, model registries, monitoring (drift, performance, bias), A/B testing, and rollback strategies
- Cloud AI on Azure (preferred): Azure AI Foundry, Azure OpenAI, Azure AI Search, Azure ML, Azure Key Vault, with solid understanding of IAM, secrets, and encryption
- Demonstrated security, privacy, and compliance competence: HIPAA, PHI/PII handling and de-identification (e.g., Presidio), Responsible AI practices
- Excellent communication and cross-functional collaboration, including with clinical stakeholders and compliance teams
- Experience with FHIR and HL7 data standards; clinical NLP (entity extraction, summarization, coding/RCM use cases); and/or medical imaging (DICOM)
- Databricks (Delta Lake, Spark), Airflow (or similar orchestration), and Azure-native data services (e.g., Data Factory, Synapse)
- GPU/CUDA experience; inference optimization (quantization, distillation); prompt/token budgeting and caching strategies for LLM workloads
- Governance & ethics: model cards, datasheets for datasets, bias/fairness evaluations, and red-teaming
- Familiarity with .NET microservices and API design to integrate AI services into enterprise systems