Role & responsibilities

Agentic AI Development Project Brief Project description: The system utilizes specialized LLM agentsorchestrated via an Agent Development Kit (ADK)to perform automated data retrieval, multi-step logical reasoning, and hypothesis validation. This platform integrates 'Human-in-the-Loop' protocols to ensure that autonomous actions remain aligned with organizational safety and accuracy standards, transforming how the enterprise interacts with complex data silos.

Role: AI Engineer specializing in Agentic AI Responsibilities Design and implement sophisticated Multi-Agent Systems (MAS) using frameworks like the Agent Development Kit (ADK) to solve complex, multi-step reasoning tasks. Architect "Human-in-the-Loop" (HITL) workflows, ensuring seamless hand-offs between autonomous agents and human subject matter experts for critical decision-making. Implement advanced reasoning patterns such as Chain-of-Thought (CoT), ReAct, and Self-Reflection to improve the reliability and transparency of agentic outputs.

Optimize long-context window utilization for analyzing vast datasets, ensuring agents maintain coherence and accuracy when processing millions of tokens. Establish rigorous AI Evaluation (Eval) frameworks to measure agent performance, including trajectory analysis, precision, recall, and faithfulness of reasoning. Must-Have Skills Solid understanding of Python programming Experience in GCP projects and Vertex ai is must.

AI evaluation methods/metrics. ( like Precision, Recall, and F1 score for agent trajectories and defect classification accuracy.)



Experience with Agentic AI, managing sequential and loop-based workflows. specifically using Google ADK or similar. Familiarity LLM Tuning (SFT, COT, ReAct) Experience in Data Analysis and process with multi agent architectures. Ability to debug and perform Root Cause Analysis (RCA) for issues.

Ability to deploy agents. Familiarity for implementing "Interrupt & Resume" patterns for human expert validation. Strong communication and teamwork skills, especially in the context of customer collaboration.

Positive-to-Have Skills Excellent problem-solving and analytical skills related to AI integration challenges specific to customer needs.

Experience in A2A, MCP for multi agent interaction. CICD pipeline for Multi Agents framework Experience integrating LLMs with Knowledge Graphs Using a memory bank for agent to agent communication.

Qualifications and Prior Experience 4-6+ years of experience in AI development, preferably on Google Cloud Platform.

Experience on building Agents, using Google ADK.

Experience in developing and deploying AI Solutions within an agent-based framework/Application Experience/Familiarity with Agile development.

Certifications and Trainings: Professional Machine Learning Engineer.

Google Skill

Boost labs/course Getting started with ADK: https://www.skills.google/catalog_lab/32017 Deployment: https://www.skills.google/catalog_lab/32530 Build and deploy multi agent: https://www.skills.google/course_templates/1445 Tune Gemini Model SFT: https://www.skills.google/catalog_lab/31712

📌 Agentic AI Development (Bengaluru)
🏢 Cerebra
📍 Bengaluru