Cloud Contact Center Engineer

I build AI-powered contact centers for enterprises.

Currently shipping Amazon Connect IVRs

Amazon Connect · Amazon Lex · GenAI · AWS

Get in touchView work
Portrait of Kartik DuaAmazon ConnectGenAIAWS certified ×2Gurugram, IN

0.0+ yrs

Contact centers

0%

Faster resolution

0+

GenAI environments

01 · About

Telephony meets GenAI

I'm a cloud contact center engineer with 3.5+ years building Amazon Connect solutions for banking and healthcare enterprises. My work sits where telephony meets GenAI: IVRs that let callers help themselves, Lex bots, RAG pipelines, and agent workflows that cut call volumes and handle times. I care about shipping things that survive production, not just demos.

Most people's worst experience with a company is its phone line. I build systems that fix that.

Off the clock I'm a tech enthusiast with a weakness for new gadgets, and I design websites for the fun of it. This one included.

02 · Experience

Where I've worked

  1. Tata Consultancy Services

    2026 - Present

    Amazon Connect Developer

    Gurugram, India · Client: a leading bank

    • Building and extending Amazon Connect contact center solutions for a leading bank, spanning contact flows, routing, and AWS integrations.

  2. Persistent Systems

    2024 - 2026

    Senior Software Engineer

    Client: IBM · Promoted in 2024

    • Deployed a Policy Bot using IBM Watson Assistant and Generative AI that resolved routine support queries on its own, cutting ticket resolution time by 30%.
    • Built 8+ GenAI contact center demo environments for IBM's enterprise sales pipeline, which helped close multiple enterprise deals in healthcare and banking.
    • Connected Salesforce Service Cloud to Amazon Connect through the CTI Adapter, giving agents an instant screen pop on every inbound call with cases created automatically by Lambda.
    • Built production RAG pipelines with LangChain that combine vector search with LLM generation, so answers come from internal knowledge bases rather than model guesswork.
    • Engineered Agentic AI workflows with LangChain agents and IBM watsonx that classify tickets, route them to the right place, and walk users through routine resolutions.

  3. Persistent Systems

    2022 - 2024

    Software Engineer

    Client: IBM

    • Architected Amazon Connect contact flows with Amazon Lex V2 and AWS Lambda for banking and healthcare clients, so routine queries stopped reaching live agents.
    • Implemented omnichannel routing across voice, chat, and tasks, with routing by agent skills and priority queues, and fed Contact Lens through Kinesis for call analytics.
    • Built a Support Assist Dashboard that improved ticket visibility across teams, reducing handling delays by 20%.
    • Containerized microservices with Docker and deployed them on AWS EKS, giving every environment the same standard deployment.

  4. Persistent Systems

    2022

    Software Engineering Intern

    Client: IBM

    • Set up and tested Amazon Connect instances, IVR flows, and Lambda integrations, and validated contact flow logic and CloudWatch logging with the development team.

03 · Selected Work

Things I've shipped

IBM · Production

GenAI Policy Bot

30% faster ticket resolution

A support bot built on IBM Watson Assistant and Generative AI. It answers routine policy questions on its own, pulls every answer straight from the actual policy documents so nothing is invented, and passes the conversation to a human agent the moment it is unsure.

User queryWatson AssistantGenAIPolicy docsResolvedAgentEscalation
Watson AssistantGenerative AINLPPython
Enterprise · Production

Salesforce CTI Integration

Screen pop + automated case creation

Amazon Connect wired into Salesforce Service Cloud through the CTI Adapter. When a call comes in, the agent instantly sees who is calling and their history, Lambda opens a case in the background, and updates land on the ticket while the call is still going. No agent starts a call blind.

Inbound callAmazon ConnectCTI AdapterScreen popAWS LambdaFlow eventsAuto case creation
Amazon ConnectCTI AdapterAWS LambdaSalesforce
IBM · Production

RAG Knowledge Pipeline

Document-grounded answers, instantly

Knowledge pipelines built with LangChain. Internal documents are converted into embeddings and stored in a vector database, a retriever pulls the most relevant passages for each question, and the LLM writes its answer from those passages instead of guessing. Engineers use it daily through a Streamlit interface.

DocumentsEmbeddingsVector storeQueryRetrieverLLMAnswer
LangChainVector SearchLLMsStreamlit
FreelanceVisit ↗

NilcoIndia.in

Live production website

A production website for a leather goods manufacturer, built and deployed by a team of two during college. I handled the front end and the deployment myself, and the site is still live today.

Front-endDeployment

04 · Skills

What I work with

Contact Centermy specialty

Amazon ConnectAmazon Lex V2Contact FlowsIVR DesignCTI IntegrationOmnichannel RoutingContact LensAmazon PollyAmazon Transcribe

AI & GenAI

LangChainRAG PipelinesAgentic AIAmazon BedrockPrompt EngineeringIBM watsonxLLM Integration

AWS

LambdaS3DynamoDBAPI GatewayCloudWatchIAMEKSEC2Kinesis

Languages & Frameworks

PythonSQLFlaskStreamlitBash

Tools

DockerSalesforce Service CloudGitCI/CDREST APIs

Certifications

  • AWS Certified Developer Associate2026
  • AWS Certified Cloud Practitioner2024
  • IBM watsonx Essentials

Education

Chitkara University

B.E. Computer Science & Engineering

Specialization in Data Science

2018 - 2022

05 · Contact

Let's talk

Building a contact center, exploring GenAI for CX, or hiring? My inbox is open.

kartikdua.official@gmail.com

Gurugram, India

LinkedInX