Generative AI ARCHITECT

Position: AI Architect

Location: Marysville, Ohio (Onsite 3 days a week)

Design multilayered AI solutions balancing compute efficiency, contextual fidelity, and algorithmic adaptability for retrieval, reasoning, planning, and tool use.

Develop advanced RAG pipelines leveraging vector databases (e.g. Chroma

DB, Milvus, FAISS) and embedding strategies for contextual accuracy.

Integrate AI capabilities with enterprise systems via REST and Graph

QL APIs, ensuring secure and scalable interoperability.

Establish best practices for algorithm selection and layering, combining neural models, symbolic reasoning, and tool‑based agents for optimal performance.

Collaborate with cross‑functional teams to embed AI agents into business workflows, aligning with compliance and governance standards.

Implement structured output validation and schema enforcement using Pydantic, FastAPI, and JSON Schema for robust data integrity.

Optimize compute resources and latency trade‑offs across cloud, hybrid, and edge environments for high‑performance AI workloads.

Define observability baselines, telemetry, tracing, evaluation metrics, and rollout strategies for safe iterative deployments.

• Lang Chain ecosystem:
  Lang Chain, Lang Graph, prompt templates, agent orchestration patterns
• AIML frameworks:
  OpenAI API, Hugging Face Transformers, Tensor Flow, PyTorch – experience with fine tuning and inference optimization
• Data context management: SQL Alchemy, Postgre
  
  SQL, JSON Schema mapping, feature engineering, contextual pipelines
• Vector databases & semantic search:
  Chroma
  
  DB, Milvus, FAISS, embedding optimization, similarity search strategies
• Algorithmic design: strong understanding of algorithm layering, retrieval, reasoning, planning, hybrid AI approaches, compute‑aware model selection
• API integration & security: REST, Graph
  
  QL, OAuth, enterprise‑grade security practices
• Dev Ops & CI/CD:
  Git, Docker, Azure Dev Ops (or equivalent), FastAPI, Uvicorn, containerized deployments, automated pipelines
• Agentic capabilities: reasoning adaptation, tool calling, Multi‑Agent coordination (ReAct agents, Supervisor)
• Performance optimization: distributed compute strategies, GPU/T‑GPU utilization, quantization, pruning, distillation, caching, batching

• Architectural diagrams