- Company Name
- Persistent Systems
- Job Title
- AI Architect
- Job Description
-
**Job Title:** AI Architect
**Role Summary:**
Designs and leads the enterprise architecture for next‑generation AI solutions, focusing on agentic intelligence, Retrieval‑Augmented Generation (RAG), multimodal model integration, and secure orchestration. Drives standards for model inference, safety, tool calling, governance, and end‑to‑end AI application delivery across business and engineering teams.
**Expectations:**
- 12+ years of experience in AI/ML architecture and enterprise AI deployments.
- Proven expertise in large language models (LLMs), vision‑LLMs, diffusion models, and agentic systems.
- Demonstrated ability to architect and optimize secure inference pipelines (vLLM, TGI, Hugging Face pipelines).
- Strong background in data platform technologies (Databricks, Unity Catalog, MLflow, vector search).
- Experience with document ingestion, OCR, and structured knowledge extraction.
- Deep knowledge of AI safety, guardrail frameworks, and responsible AI compliance.
**Key Responsibilities:**
- Define enterprise AI architecture incorporating Agentic RAG, Graph RAG, and multi‑agent orchestration.
- Establish guardrails for tool schemas, action validation, lineage tracking, and auditability.
- Lead end‑to‑end design of intelligent copilots, autonomous agents, workflow systems, and knowledge‑grounded assistants.
- Integrate and optimize open‑source/commercial LLMs, vision‑LLMs, and diffusion models; ensure secure, high‑performance inference.
- Design scalable knowledge‑base ingestion pipelines for PPT/PDF, including OCR, layout parsing, semantic chunking, and metadata enrichment.
- Develop safety, grounding, and evaluation frameworks using DeepTeam, NVIDIA NeMo Guardrails, and related tools.
- Collaborate with data, security, and product teams to enforce responsible AI practices and compliance.
- Drive continuous improvement of AIOps and AgentOps patterns, tooling, and governance standards.
**Required Skills:**
- AI/ML architecture, agentic AI, RAG / graph RAG, multimodal models (vision‑LLMs, diffusion).
- Inference optimization (vLLM, TGI, Hugging Face pipelines).
- Multi‑agent orchestration, tool calling, safe action schemas.
- Secure API integration, enterprise security frameworks.
- Databricks (MLflow, Unity Catalog, Vector Search), Databricks‑first pipelines.
- Document ingestion & OCR (Unstructured.io, Tika, PyMuPDF, LayoutParser, docTR, PaddleOCR).
- AI safety governance, guardrail frameworks (DeepTeam, NVIDIA NeMo).
- Evaluation and compliance toolsets.
**Required Education & Certifications:**
- Bachelor’s or Master’s degree in Computer Science, Engineering, Data Science, or related field.
- Professional certifications in AI/ML (e.g., TensorFlow, PyTorch, Databricks Engineer).
- Preference for advanced certifications in responsible AI, AI governance, or data platform expertise.