You will get A Working, Production-Ready App Fix
Rising Talent
Rising Talent
Project details
You built your app with Lovable, Cursor, Bolt, or another AI tool and now something's broken. Deployment fails. The chatbot hallucinates. Auth won't connect. You're stuck.
I fix AI-built and half-finished apps. Same-day turnaround for most bugs. I diagnose the root cause, fix it in your repo via pull request, test in the deployed environment, and give you a written summary of what changed and why.
I work with TypeScript, Python, React, Next.js, FastAPI, Claude API, OpenAI, Supabase, PostgreSQL, Docker, Vercel, and Azure. Every fix is deployed, tested, and verified. Fixed price, no surprises.
I fix AI-built and half-finished apps. Same-day turnaround for most bugs. I diagnose the root cause, fix it in your repo via pull request, test in the deployed environment, and give you a written summary of what changed and why.
I work with TypeScript, Python, React, Next.js, FastAPI, Claude API, OpenAI, Supabase, PostgreSQL, Docker, Vercel, and Azure. Every fix is deployed, tested, and verified. Fixed price, no surprises.
AI Development Type
Knowledge Representation, Software MaintenanceAI Development Language
PythonWhat's included $200
These options are included with the project scope.
$200
- Delivery Time 1 day
- Number of Revisions 1
- Detailed Code Comments
- Source Code
Optional add-ons
You can add these on the next page.
Additional Revision
+$50Frequently asked questions
1 review
(1)
(0)
(0)
(0)
(0)
This project doesn't have any reviews.
CM
Christian M.
Oct 12, 2022
Write Programming or Web Related Tutorial
MacDara is a very skilled freelancer and web developer. Highly recommended!
About MacDara
AI Engineer | Claude Agents, MCP, RAG, Evals | Prod-Grade LLM Systems
Claremont, Australia - 2:49 pm local time
Not "AI-curious." I ship.
Stack I work in daily:
- Claude (Anthropic API), OpenAI, Gemini - model selection, prompt engineering, cost optimisation
- Claude Code, Claude Agent SDK, MCP servers (Model Context Protocol)
- Multi-agent orchestration, tool use, structured outputs
- RAG with semantic search, vector databases, hybrid retrieval
- Eval pipelines - promptfoo, custom LLM-judge, regression suites
- FastAPI + Python backends, Next.js + TypeScript frontends
- Postgres, Redis, Railway, Vercel, Docker, Azure
Three live platforms you can poke at:
1. AzurePrep - paid cloud certification platform covering 35 certs and 15,000+ questions. Stripe payments, adaptive study flows, drag-drop interactions, full E2E test suite.
2. Chlann - messaging platform with ML-powered photo search and mobile clients. Async indexing, performance tuning, Capacitor cross-platform builds.
3. wuts.press - forensic OSINT investigation site. Automated data pipelines, AI-assisted publishing, government/regulatory source integrations, custom search and citation tooling.
Also currently building CS Pulse - a multi-tenant GTM intelligence platform for B2B marketing agencies, with HubSpot / Lusha / Instantly integrations and a Claude-powered analyst chat with guardrails.
How I work:
- Plan before building. Scoping calls, written architecture, eval criteria - all before code
- I write tests. Real tests, not theatre
- I report honestly. If a build is at risk, you'll hear it from me before the deadline
- I don't disappear. Daily updates on active engagements
Best fit:
- Founders shipping their first AI product and want to avoid expensive mistakes
- Technical teams who need a senior pair-of-eyes on prompt / agent architecture
- Companies with a working MVP that's slow, expensive, or unreliable in production
Not a fit:
- "Build me ChatGPT but for X" with no scoping
- Pure prompt-tweaking with no engineering
- Sub-$500 fixed-price projects
Based in Perth, Australia (UTC+8). I work async-friendly with US, EU, and APAC clients - my hours overlap morning APAC, evening US, and full EU business day.
Start with a 30-min consultation if you want to scope a project, or message directly if you've already got a written brief.
Steps for completing your project
After purchasing the project, send requirements so MacDara can start the project.
Delivery time starts when MacDara receives requirements from you.
MacDara works on your project following the steps below.
Revisions may occur after the delivery date.
Diagnose
I review your codebase, reproduce the bug, and identify the root cause. You get a summary of what's wrong before I change anything.
Fix and test
I implement the fix, test it locally and in the deployed environment, and submit via pull request or direct commit.