TIME WAIT BLOG.
Vol. 2024.10

" Time Wait in .cc — Slow down for better code. "

Harry Shao

An open engineering journal documenting systems, performance, and pragmatic product delivery.

/latest_thoughts

42 entries found
100 Q&As on FDE: From Career Transitions and Solo Freelancing to Enterprise AI Delivery
#AI Tutorials • Oct 16, 2026

100 Q&As on FDE: From Career Transitions and Solo Freelancing to Enterprise AI Delivery

Being a Forward Deployed Engineer (FDE) is a role, but shipping a technical project and building a sustainable enterprise AI delivery business are two entirely different disciplines. This 23,000-word handbook breaks down 100 concrete, battle-tested questions covering career transitions, skill gaps, portfolio construction, client discovery, technical trade-offs (prompts vs. RAG vs. agents), quotation and milestone billing, stakeholder alignment, and production handover.

read full log arrow_right_alt
What Does a Million-Yuan FDE Project Actually Deliver?
#AI Tutorials • Oct 10, 2026

What Does a Million-Yuan FDE Project Actually Deliver?

FDE is not an abstract theoretical concept—it is about engineers entering the client's business trenches. This retrospective deconstructs the complete delivery of a seven-figure enterprise AI office system: phased rollouts, strict delivery boundary scoping, architectural trade-offs between SaaS and on-prem source code delivery (model gateways, lightweight BM25 retrieval, deterministic script safeguards), end-to-end file acceptance, and dual-ledger ROI accounting.

read full log arrow_right_alt
The First Step in FDE: How to Clean Massive Multimodal Data into a Trustworthy Dataset (A Technical Deep Dive)
#AI Tutorials • Oct 04, 2026

The First Step in FDE: How to Clean Massive Multimodal Data into a Trustworthy Dataset (A Technical Deep Dive)

Cleaning massive multimodal data is far more than chaining model inferences. This deep dive systematically deconstructs enterprise data cleaning across text, image, audio, and video: three-tier storage with evidence pointers, bounding box attribution, ASR-video alignment tolerances, log-odds conflict resolution, multi-stage deduplication, anomaly triage, and golden dataset regression.

read full log arrow_right_alt
How an Air Fryer DTC Brand Scaled Daily Orders from 30 to 210 via Reddit and AI Search
#Reddit Operations • Sep 30, 2026

How an Air Fryer DTC Brand Scaled Daily Orders from 30 to 210 via Reddit and AI Search

Skipping paid ads and scaling daily orders from 30 to nearly 210? Here is the full breakdown of how an air fryer indie e-commerce store cracked overseas organic growth: identifying high-intent community debates, crafting AI-citation-friendly answers with Claude, transforming product landing pages, and riding the Reddit x AI Search (GEO) flywheel.

read full log arrow_right_alt
Why Fake Self-Promotion on Reddit Gets Caught Immediately (And How to Do It Right)
#Reddit Operations • Sep 23, 2026

Why Fake Self-Promotion on Reddit Gets Caught Immediately (And How to Do It Right)

Buying high-karma accounts only to get banned in minutes? Astroturfing with alt accounts only to get publicly mocked? On Reddit, fake grassroots marketing is suicide. Here is why overseas users spot sockpuppets instantly, and how indie hackers can build genuine credibility, avoid shadowbans, and acquire real users.

read full log arrow_right_alt