DK
Back to Work
AI Product·May 2024 – Present

The Branding Club

Head of Engineering & AI Architect

100K/sRead/write ops sustained
100KSKUs generated per batch
70%Manual content effort cut
20Engineers hired & led

Overview

The Branding Club is a brand content and SKU generation SaaS. I joined as the Head of Engineering and built the company's entire technical foundation — establishing a 20-member cross-functional team (Engineering, QA, DevOps) from zero and architecting Brandhub, the core product: a high-throughput microservices platform capable of generating 100,000 SKUs per batch with AI-powered content at its core.

The Challenge

The product needed to process massive volumes of product data and generate high-quality brand content at scale — 100,000 SKUs per batch — without sacrificing speed or consistency. A naive LLM-per-request approach would be too slow and too expensive. The architecture needed multiprocessing, intelligent queuing, and an LLM evaluation layer to maintain output quality across all AI workloads.

What I Built

01

Established and scaled a 20-member cross-functional team (Engineering, QA, DevOps) from zero — driving hiring pipelines, engineering culture, onboarding, and technical standards.

02

Architected Brandhub's core microservices platform: multiprocessing workers, multithreaded queuing, and database sharding to sustain 100,000 read/write ops/second and generate 100,000 SKUs per batch.

03

Designed the AI content and SKU generation layer using RAG pipelines and LLM APIs (GPT-4, Claude) — cutting manual content production effort by 70% vs the previous workflow.

04

Introduced LLM evaluation and observability tooling (Helicone, LangSmith) to monitor output quality, reduce hallucinations, and optimise token costs across all AI workloads.

05

Built the multi-tenant SaaS architecture on Django — per-tenant data isolation, JWT-scoped API access, Stripe subscription billing, and a role-based admin portal.

06

Deployed on AWS (EC2 Auto Scaling, RDS PostgreSQL, ElastiCache Redis, S3, CloudFront) with Kubernetes orchestration and GitHub Actions CI/CD.

Tech Stack

AI / LLM

GPT-4Claude APILangChainRAG PipelinesHeliconeLangSmith

Backend

DjangoDRFCeleryRedisPostgreSQLKafka

Frontend

Next.jsTypeScriptTailwind CSSRecharts

Infrastructure

AWS EC2KubernetesRDSElastiCacheS3CloudFront
Next ProjectCeodekho