Vishwanath ReddySoftware Engineer II
All projects

May — Dec 2024

EngageON & Tellow AI

Two SaaS platforms with event-driven backends, AI image and video generation, character training, and integrated product dashboards.

FocusFull-stack product development

StatusProfessional product work

  • Next.js
  • Node.js
  • Redis
  • MySQL
  • ClickHouse
  • Kafka
  • Cloudflare R2
  • Razorpay

Overview

Across EngageON and Tellow AI, the work spanned product interfaces, backend systems, AI model integration, admin tools and payments. The architecture combined Next.js and Node.js with Redis, MySQL, ClickHouse and Kafka.

The problem

AI products need to connect changing model capabilities to usable customer workflows, while keeping asynchronous product events, dashboards and payments reliable.

Implementation

  • Built two SaaS platforms, EngageON and Tellow AI, using Next.js, Node.js and Redis, MySQL and ClickHouse.
  • Designed scalable backend systems and Kafka-based event-driven workflows to reduce API response times.
  • Integrated image and video generation and character-training pipelines with flexible model integration.
  • Built reusable backend and frontend components that reduced AI model onboarding time by 50%.
  • Developed dashboards, admin systems and user panels, and integrated Razorpay payments.

Outcome & status

Reusable components reduced AI model onboarding time by 50%. The source details provided no before-and-after API latency figures, so this page describes the improvement without an invented number.

From input to outcome

How it works

  1. Build product surfaces

    Next.js interfaces provide end-to-end dashboards, admin tools and user panels.

  2. Orchestrate workflows

    Node.js services coordinate product operations and Kafka event-driven workflows.

  3. Integrate AI models

    Reusable backend and frontend components connect image/video generation and character-training pipelines.

  4. Store product data

    Redis, MySQL and ClickHouse support the multi-database architecture.

Inside the implementation

Engineering decisions

Keep model integration reusable

Shared frontend and backend components let new AI models join product workflows without rebuilding each integration, cutting onboarding time by 50%.

Use events for asynchronous work

Kafka-based workflows decouple product events from request handling and were used to reduce API response times.

Connect admin, user and payment flows

Dashboards, administrative tools, user panels and Razorpay integration were built as parts of the same product experience.

Interested in how I build?

Let’s talk