Agentic AI Architect

Job Description:

• Architect and implement agentic GenAI systems using modern LLM orchestration frameworks (LangChain, LangGraph, CrewAI, AutoGen, etc.).
• Design multi-agent architectures including planner–solver patterns, hierarchical agents, autonomous teams, and goal decomposition workflows.
• Define end-to-end GenAI architecture covering model selection, orchestration, memory, tools, security, observability, and deployment.
• Build reusable agent tooling, APIs, and memory architectures for coordination, context persistence, and scalable interaction.
• Lead hands-on development and production deployment of GenAI applications such as assistants, copilots, workflow automation, and decision-support systems.
• Evaluate, integrate, and optimize LLMs and multimodal models (OpenAI, Claude, Gemini, Mistral, LLaMA, etc.) and RAG pipelines using vector databases and retrieval systems.
• Optimize systems for performance, reliability, scalability, cost efficiency, and responsible AI constraints (security, privacy, bias, hallucinations).
• Establish and enforce GenAIOps / LLMOps practices including evaluation, monitoring, logging, feedback loops, and continuous improvement.
• Guide and mentor teams on agent design patterns, prompt engineering, tool abstraction, fine-tuning, and best practices.
• Partner with product, business, and domain stakeholders to translate business requirements into scalable agent-based solutions and AI roadmaps.
• Drive technology strategy and innovation by tracking advances in multi-agent systems, cognitive architectures, and open-source AI frameworks.

Must have skillsets:

• 10 years of overall experience, including 5+ years in Pharma and Biotech, with a strong track record in building domain-specific GenAI agents for Life Sciences use cases.
• Hands-on expertise in agentic and autonomous AI systems, including multi-agent architectures and frameworks such as LangChain, LangGraph, CrewAI, AutoGen, Haystack, AutoGPT, Baby AGI, CAMEL, and MetaGPT.
• Strong Python proficiency with deep experience in prompt engineering, tool integration, agent orchestration, RAG pipelines, vector databases, and memory architecture.
• Practical experience working with commercial and open-source LLMs (GPT-4, Claude, Gemini, Mistral) and implementing fine-tuning, adapters, and RLHF where applicable.
• Proven ability to design, deploy, and operate cloud-native GenAI solutions on AWS, Google Cloud Platform, and Azure, with focus on latency optimization, autoscaling, and cost control.
• Demonstrated leadership in architecture design, proof-of-concept development, and hands-on delivery in fast-paced environments.
• Experience with model cost profiling, real-time performance tuning, and scalability strategies, ensuring reliable and efficient agent behaviour
• Working knowledge of agent simulation, environment modelling, or reinforcement learning is a plus, along with strong awareness of compliance, privacy, and safety requirements for enterprise GenAI deployments.