Compare
Fashionly vs Joel Transport vs Makunike.com AI Twin
The problem, what was built, the stack and the outcome, next to each other.
Clear · Up to 3 at a time
In common
- Shared stack: none in common
- Both kinds of work: Web
- Order: Joel Transport (Dec 2022 – May 2023) → Fashionly (Apr 2026 – present) → Makunike.com AI Twin (Oct 2026)
Swipe sideways to compare ›
Fashionly ›
Your style, elevated: an AI stylist that starts with what you own
Joel Transport ›
Online booking platform on AWS
Makunike.com AI Twin ›
This website: a portfolio you can interview by phone call or chat
At a glance
- Role
- Founder / Lead Engineer
- When
- Apr 2026 – present
- Length
- 7 mo so far
- Status
- MVP
- Type
- AI, Mobile, Web
- Role
- Lead Engineer
- When
- Dec 2022 – May 2023
- Length
- 6 mo
- Status
- Live
- Type
- Web, Platform
- Role
- Designer & Engineer
- When
- Oct 2026
- Length
- 1 mo
- Status
- Live
- Type
- AI, Web
The problem
Fashion-commerce recommendations usually optimise for another purchase without understanding what the customer already owns.
Centralised the customer journey and operational data, reduced manual hand-offs and set up a documented AWS foundation.
Portfolios are static, but hiring conversations aren't. A recruiter wants to know whether I've shipped mobile apps, a hiring manager wants a production story in STAR form, and a sourcing agent wants structured answers it can compare. A PDF CV can't do any of that.
What was built
Fashionly grounds recommendations in the user's wardrobe, builds outfits from existing items and recommends products only where a genuine wardrobe gap exists. Khanya, a LangGraph stylist with about 15 typed tools, handles wardrobe access, catalogue search, saved looks and human-stylist bookings.
Led architecture and hands-on engineering for an online booking platform: Python APIs and the AWS deployment architecture, integrated with Paystack, Sage Accounting, Zoho Bigin and Zoho SalesIQ.
A digital version of me that holds a real interview by phone-style voice call or text, grounded only in material I curated. It changes what's on screen as it answers, logs what it can't answer so I can teach it, records and QA-scores every call, and is reachable by other AI agents over MCP and A2A.
Stack
- Next.js
- Kotlin
- Supabase
- LangGraph
- MongoDB
- SSE
- Python
- AWS
- Paystack
- Sage
- Zoho
- Next.js 15
- React 19
- Tailwind CSS v4
- GPT-6.1 (Azure, Responses API)
- gpt-realtime-2.1
- WebRTC
- text-embedding-3-small
- Supabase pgvector
- TypeSafe Jev
- MCP
- A2A
- Azure Blob
- LangSmith
- Vercel
Highlights
- Khanya: a LangGraph tool-calling stylist with about 15 typed tools
- MongoDB-checkpointed conversation memory and SSE streaming of tool progress and grounded product cards
- Request-scoped Supabase clients with row-level security on every request
- Python APIs and AWS deployment architecture
- Paystack, Sage Accounting, Zoho Bigin and Zoho SalesIQ integrations
- Generative UI: the model calls typed navigate and show_panel tools, streamed to the browser as NDJSON events
- Phone-call voice on gpt-realtime-2.1 over WebRTC, with every call recorded, transcribed and QA-scored
- TypeSafe Jev in ten places: routing, tool selection, interview mode, guards, lead scoring, grounding, triage, FAQ review, agent vetting and call QA
- Self-improving RAG: weak or ungrounded answers land in a ranked CMS inbox and become FAQs
- MCP and A2A servers with request vetting and rate limits for agent-to-agent interviews
Infrastructure choices
- Supabase with RLS
- MongoDB checkpoints
- SSE
- AWS
- Vercel serverless (Next.js route handlers)
- Azure OpenAI
- Supabase Postgres + pgvector (HNSW)
- TypeSafe Jev
- Azure Blob Storage
Outcome
A working end-to-end MVP across web and Android: wardrobe management, catalogue discovery, conversational styling, saved looks and human-stylist booking.
Centralised the customer journey and key operational data, reduced manual hand-offs and established a documented AWS foundation for future development.
The twin never invents a fact about my career: when it doesn't know, it says so and I'm notified. Intent-matched navigation fires within about a second, before the first answer token. Every call is recorded, transcribed and scored so review starts with the calls that need it, and one knowledge base serves text, voice, MCP and A2A.