Key Responsibilities

1. Enterprise AI & GenAI Architecture

● End-to-End System Design: Architect scalable, low-latency, and highly available AI/ML systems that seamlessly integrate Classical ML (predictive modeling, clustering, time-series) with Generative AI (LLMs, multimodal models).

● Agentic Frameworks & Multi-Agent Systems: Design autonomous and semi-autonomous AI agents using modern frameworks (e.g., LangGraph, AutoGen, CrewAI). Architect complex agentic reasoning workflows involving tool-use, multi-step planning, and dynamic execution.

● State Management & Memory: Engineer advanced conversational memory architectures (short-term, long-term, semantic memory) and robust state management protocols for long-running agentic tasks.

● Advanced RAG & Hallucination Mitigation: Build highly accurate RAG pipelines utilizing vector databases (Pinecone, Milvus, Weaviate), hybrid search algorithms, and semantic reranking. Implement rigorous guardrails, self-correction loops, and grounding techniques to systematically eliminate LLM hallucinations.

2. Cloud & Modern Data Platforms

● Multi-Cloud AI Infrastructure: Lead architecture decisions and deployment strategies across AWS (Bedrock, SageMaker, EKS), Azure (OpenAI, ML Studio, AI Search), and GCP (Vertex AI, BigQuery ML).

● Up-to-date Data Architecture: Solve complex data engineering challenges, including streaming vs. batch ingestion, unstructured data processing, and scalable embedding generation. Architect solutions natively integrated with modern data platforms (e.g., Snowflake, Databricks, Apache Kafka).

● Data Mesh & Governance: Define data contracts and implement architectures that ensure clean, compliant, and governed data pipelines to feed AI models in real-time.

3. Production-Grade Engineering & LLMOps

● Scalability & Resilience: Translate PoCs into robust, scalable, and fault-tolerant production systems. Establish architectural patterns for high-throughput model serving, caching, API gateways,



and load balancing.

● MLOps & LLMOps: Establish enterprise standards for CI/CD in machine learning. Oversee automated model evaluation, continuous fine-tuning (LoRA, PEFT), prompt lifecycle management, drift detection, and system observability.

- Security, Compliance, & Responsible AI

● Security by Design: Embed zero-trust architecture, data encryption (at rest and in transit), Role-Based Access Control (RBAC), and Data Loss Prevention (DLP) natively into the AI ecosystem.

● AI Governance: Ensure all AI models and systems comply with global regulatory standards (ISO/IEC 42001, NIST AI RMF, EU AI Act, SOC2, HIPAA).

● Ethical AI: Implement bias detection mechanisms, explainability protocols (XAI), and human-in-the-loop (HITL) fallback strategies.

Required Qualifications & Technical Expertise

● Experience: 12+ years of software engineering and architecture experience, with a minimum of 5+ years dedicated to AI/ML architecture and deploying models into production environments.

● Generative AI & LLMs: Deep understanding of foundational models (GPT, Claude, Llama , Gemini). Hands-on experience with fine-tuning methodologies, prompt engineering, and embedding generation.

● Agentic & Orchestration Frameworks: Master-level proficiency in LangChain, LlamaIndex, LangGraph, AutoGen, or equivalent ecosystem tools.

● Cloud Native Mastery: Extensive architectural experience across AWS, Azure, and/or GCP. Strong command over containerization and orchestration (Docker, Kubernetes/EKS/AKS/GKE).

● Modern Data Stack: Proficiency with vector databases, graph databases (Neo4j), and modern data warehousing/lakehouse architectures (Databricks, Snowflake).

● Programming Languages: Expert-level Python; strong proficiency in Go, Java, or C++ is highly preferred.

● Leadership & Vision: Exceptional ability to communicate complex architectural tradeoffs to C-level executives, while also being capable of doing deep-dive code and architecture reviews with senior engineering teams.

📌 Senior Principal AI/ML Architect (Pune)
🏢 CJET Tech Solutions
📍 Pune