Tan LE

  Post 3 in my series on an automation hub for a ticket-resale business. If you've ever put an LLM into a production pipeline, you know the feeling: the prompt clearly says "return ONLY valid JSON," and the model responds with... a friendly greeting, a markdown code fence, and then the JSON — occasionally missing a bracket. This post collects the techniques I use to parse LLM output safely in C#. The problem: an LLM is an "API" with no schema guarantee In my email-classification pipeline (see post 2), the model must return a list of objects like: { "uid": "452455", "subject": "Presale starts Friday!", "performer": "...", "venue": "...", "event_date": "Oct 10, 2026", "label": "PRESALES", "confidence": 0.92 } In practice the model returns every possible variation: uid is sometimes a string and sometimes a number, label i...

This is a living index of the systems I'm most proud of — each one is in daily production use, and each links to a deeper technical write-up. If you're a recruiter or hiring manager, this page is the fastest way to see what I actually build. 1. AI Email-Triage Pipeline (LLM classification in production) A pipeline that classifies hundreds of ticketing emails per day (presales, releases, cancellations) using multimodal LLMs, then distributes summarized reports automatically to the relevant business departments. I fine-tuned GPT-4o-mini on a labelled dataset I built from PDF archives, put OpenAI and Gemini behind a single interchangeable JSON contract for cost/performance trade-offs, and wrote a defensive C# parsing layer — balanced-brace JSON extraction, tolerant custom JsonConverters, code-enforced business invariants — so a single malformed model response never breaks the batch. Stack: C#/.NET, OpenAI API (fine-tuning), Gemini, IMAP, SQL Server Read more: Classifying E...

  Post 2 in my series on an automation hub for a ticket-resale business. This one covers the AI pipeline that reads and classifies event-ticket emails — from prompt engineering and multimodal handling (emails with images) to building my own dataset and fine-tuning GPT-4o-mini. The problem The purchasing team's inbox receives hundreds of emails a day: Ticketmaster presale announcements, venues releasing new ticket blocks, artists sending promo codes, cancelled orders that need refunds... Missing a "tickets released" email can mean losing a chance to buy at face value; missing a cancellation email means losing refund money. But nobody can read hundreds of emails a day by hand. The requirement: classify emails into categories ( PRESALES , TICKETS_RELEASED , TICKETS_ON_SALE , ALMOST_SOLD_OUT , PROMOTION_DISCOUNT , Cancelled ...) and extract structured data : performer, venue, event date, purchase link, promo code — and for cancellation emails, the order number, vendor, quanti...

  This is the first post in a series about a project I built for a US-based ticket-resale company: a .NET "automation hub" that connects more than 10 different systems — from POS and marketplaces to banking APIs and an AI that reads email. The problem The event-ticket resale business in the US runs on a patchwork of disconnected platforms: SkyBox (VividSeats' POS for inventory/invoices), Lysted (a resale marketplace), Ticket Evolution, TicketUtils, SeatScouts/AutoQ (purchasing-account management), plus banking (Taekus), Slack, Gmail, and Google Sheets. Every day the operations team had to: Reconcile Lysted sales against inventory in SkyBox Record settlement payments from remit CSV files onto individual invoices Read hundreds of marketing/presale emails so no new ticket drop was missed Monitor balances and sweep cash-back into the main account Bulk-configure a fleet of purchasing accounts All of it repetitive, all of it error-prone when done by hand. My goal: fold everythi...

Earlier this year I set out to build a SaaS platform for domain name research — a tool that helps investors find newly registered domains and expiring auctions worth buying, backed by real business signals instead of gut feeling. Four months and a hundred-plus commits later, it's live. Here's what the journey actually looked like. What the platform does The core idea: a domain name is more valuable if real companies are already using that name. So the platform cross-references millions of domain candidates against company datasets — LinkedIn company data, Crunchbase, funding records — and surfaces metrics like "how many companies use this keyword" and "does the .com resolve to a funded startup." Users can filter new registrations and live auctions by dozens of these signals, save filter sets, and even schedule automatic bids on auctions. Tech stack: React 19 + Vite on the frontend (deployed to Azure Static Web Apps), ASP.NET Core 8 API on the backend, SQL Se...

 What is the webhook?  Ask Google :) 😅  This is my experience to build a Azure Function HTTP trigger to receive an event from SkyBox API Webhook.  Business requirement: Real time integration INVOICE with SkyBox API Webhook. Means whenever SkyBox has a new or Update invoice. Their service would send the json message to my API. Solution:  Step 1: Built C# HTTP trigger function and then deploy it to Azure Function Depends on the business requirement, you can add more in your own code.  Step 2:  Added a subscription sending a POST to https://skybox.vividseats.com/services/webhooks {   "topic" :  "INVOICE" ,   "url" :  "https://xxxxxxxxinvoice.azurewebsites.net/api/SkyboxReceiveInvoice?code=xxxxxxxxxx" ,   "headers" :  "Bearer: xxxxxxxx" ,   "secret" :  "yoursecretkey" } Navigate to Azure Function Monitor to see the message coming. Note: The "secret": "yoursecretkey" is using for Encrypted...