Principal Automation Developer

Percepta; Hyderabad; On-site; INR 2,000,000 - 3,200,000; Full time

Job summary

TTEC is scaling its Automation & AI portfolio to move beyond manual processes toward an agentic framework across business units. This role centers on building multi-agent systems, integrations, and evaluation rigor to deploy AI agents safely and at scale.

The position involves designing agent workflows with ADK, exposing MCP tools, and maintaining evaluation frameworks to ensure accuracy and safety across deployments.

Qualifications

• Hands-on experience designing multi-agent workflows with Google ADK or similar frameworks.
• Experience building/exposing MCP servers/tools for internal systems.
• Strong prompt engineering with evaluation-driven iteration.
• Experience with Microsoft Copilot Studio for enterprise agents/bots.
• Knowledge of GCP services for AI workloads (Vertex AI, Functions/Run, IAM).
• Solid software engineering basics: Python, API design, version control.

Responsibilities

• Design multi-agent workflows using Google ADK, including orchestration and memory management.
• Build integrations exposing internal systems as discoverable tools via MCP.
• Develop evaluation frameworks to measure accuracy, task completion, and safety.
• Develop and optimize agents in Microsoft Copilot Studio, integrating with Power Platform.
• Apply prompt engineering practices to improve reliability and reduce hallucination.

Skills

Prompt engineering

LLM evaluation methodology

Python

API design

Version control

Tools

Google ADK

LangGraph

AutoGen

CrewAI

Model Context Protocol (MCP)

Microsoft Copilot Studio

GCP Vertex AI

Cloud Functions

Cloud Run

IAM

Job description

Job Description:

About the Role

TTEC is scaling its Automation & AI portfolio to move beyond manual, ticket-by-ticket process work toward an agentic framework that can right-source work across business units. This role is central to that shift — building the multi-agent systems, integrations, and evaluation rigor needed to deploy AI agents safely and at scale across TTEcs technology and business operations.

What Youll Do

• Design and implement multi-agent workflows using Google Agent Development Kit (ADK), including agent orchestration, tool routing, and memory/context management
• Build integrations using Model Context Protocol (MCP) to expose internal systems (Jira, SharePoint, ServiceNow, internal APIs, etc.) as discoverable tools that agents can call at runtime
• Develop and maintain evaluation frameworks to measure agent accuracy, task completion rate, and safety — both pre- and post-deployment
• Design, build, and optimize conversational and task-based agents in Microsoft Copilot Studio, integrating with Power Platform (Power Automate, Dataverse) and enterprise data sources
• Apply strong prompt engineering practices — structured prompting, few-shot design, guardrails, and iterative evaluation — to improve agent reliability and reduce hallucination/drift
• Architect and deploy agentic solutions on Google Cloud Platform (GCP) (Vertex AI, Cloud Functions, Cloud Run, IAM) with attention to scalability, cost, and security
• Partner with business unit stakeholders to identify manual, repeatable processes suitable for agent-based automation, translating business requirements into technical agent designs
• Collaborate with the Automation & AI team on reusable patterns, shared tooling, and governance standards (security review readiness, access control, data handling)
• Contribute to internal enablement — documentation, training, and knowledge transfer so the broader team can build and extend agentic solutions
• Monitor deployed agents in production, iterating based on performance data, user feedback, and evaluation results

What Youll Bring

Required:

• Hands-on experience building with Google Agent Development Kit (ADK) or comparable agent orchestration frameworks (LangGraph, AutoGen, CrewAI)
• Practical experience with Model Context Protocol (MCP) — building or consuming MCP servers/tools
• Strong prompt engineering skills, including evaluation-driven prompt iteration
• Experience with Microsoft Copilot Studio for building and deploying enterprise agents/bots
• Working knowledge of GCP services relevant to AI workloads (Vertex AI, Cloud Functions/Run, IAM, logging/monitoring)
• Solid software engineering fundamentals — Python (preferred), API design, version control
• Understanding of LLM evaluation methodology: accuracy, task completion, safety/guardrail testing