Build an analysis-ready dataset
Cleaning, exception isolation and session grouping produced 578,119 cleaned records and 44,059 grouped conversations.
Conversation Analysis
Turn fragmented conversations into traceable business data.
A data processing platform and event-analysis prototype connecting record cleaning, session grouping, event extraction and source traceability.
- cleaned records
- 578,119
- grouped conversations
- 44,059
A better way to work
BEFORE
Pricing questions, order updates, after-sales issues and new requests are mixed across chat records. One conversation can contain several distinct business needs.
WITH THE PRODUCT
Clean and group historical records into traceable sessions, then use an event-analysis prototype to connect business topics with source messages for review.
Core capabilities
Make processing recoverable and traceable
Batch transactions, a resumption ledger and source links make processing progress and record origins inspectable.
Prototype event-level analysis
Explore multiple events within a conversation, including pricing, fulfillment, after-sales issues and new requests, linked to their source messages.
Evaluate quality across dimensions
Evaluate extraction, classification and evidence location separately, with human consensus samples and regression checks guiding iteration.
How it works
- 01
Collect historical records
- 02
Clean and group sessions
- 03
Extract events and link sources
- 04
Evaluate and iterate
PRODUCT / ENGINEERING
What I delivered
I audited the original system and independently designed and implemented an isolated V2 data layer, recoverable processing workflows and an event-analysis prototype. Human-reviewed samples and regression evaluation connect data engineering with business understanding and model improvement.
Does your team face a similar challenge?
Start with one workflow. Let’s explore how AI could fit your operations.
Keep exploring
All workEnterprise fulfillment assistant0→1 · Solo build
LD Assist
An enterprise fulfillment agent powered by a locally deployed large language model. Connect orders, purchasing, warehouse records and follow-ups so support teams can check progress, confirm batches and prepare reply drafts.
- support users
- 30
- queries per day
- ~300
跟一下
Personal follow-up app0→1 · Solo build
跟一下 (Genyixia)
A lightweight iOS app for replies, refunds and documents that still need follow-up. I independently designed and built the product, connected its paid features and shipped it on the App Store.
- Available now
- App Store
- Native app
- iOS
- Subscription & lifetime
- Pro