ROLE OVERVIEW:

We are seeking an experienced Forward Deployed AI/ML Engineer with 6–8 years of engineering experience to design, build, and deploy high-impact AI/ML and GenAI solutions end-to-end. This role bridges applied machine learning, software engineering, and stakeholder problem-solving to deliver production-grade systems that are scalable, secure, observable, and aligned to key business KPIs.

KEY RESPONSIBILITIES:

Use-Case Discovery & Forward Deployment: Partner with business and technical stakeholders to assess data readiness, run discovery workshops, and rapidly prototype production-grade AI opportunities.

Applied ML & GenAI Engineering: Develop classic ML models (classification, forecasting, NLP) and GenAI pipelines (RAG, document ingestion, chunking, vector retrieval, reranking, and prompt guardrails).

Productionization & MLOps/LLMOps: Package models into scalable services using Docker and Kubernetes. Establish automated CI/CD pipelines, model registries, drift detection, and monitoring.

Systems Integration & Platform Collaboration: Embed AI capabilities via REST APIs and event-driven services, ensuring compliance (PII/PHI handling), RBAC, secrets management, encryption, and audit trails.

Technical Leadership & Enablement: Lead design reviews, author architecture specifications and operational playbooks, and create reusable deployment patterns.

CORE TECHNICAL STACK

Primary Programming : Python (automation, data pipelines, backend engineering)

Frontend Stack: React.js (Primary) or Angular

Backend & API FastAPI, Node.js, REST APIs, OAuth2/JWT authentication, API Gateways (Azure API Management, Apigee)

AI/ML & GenAI Frameworks Azure OpenAI / OpenAI APIs, LangChain, Semantic Kernel, Vector Search (Pinecone, Azure AI Search, Weaviate, Milvus)

Cloud & MLOps Azure (AKS, App Services, Azure Functions, Blob Storage, Cosmos DB), AWS (Lambda, ECS/EKS, S3), Docker, Kubernetes, Databricks, MLflow

Databases SQL RDBMS, NoSQL (MongoDB, Cosmos DB)

Domain Standards HL7/FHIR (Healthcare/FDA domain exposure preferred