Skip to main content

Leverage Record: March 11, 2026

AI Time Record

About the author: I'm Charles Sieg, a cloud architect and platform engineer who builds apps, services, and infrastructure for Fortune 1000 clients through Vantalect. If your organization is rethinking its software strategy in the age of AI-assisted engineering, let's talk.

About These Records
These time records capture personal project work done with Claude Code (Anthropic) only. They do not include work done with ChatGPT (OpenAI), Gemini (Google), Grok (xAI), or other models, all of which I use extensively. Client work is also excluded, despite being primarily Claude Code. The actual total AI-assisted output for any given day is substantially higher than what appears here.

Thirty-five tasks on Tuesday. The day's defining thread was a cross-platform user profile system that touched six repositories: backend services, three client applications (web, desktop, iOS), and architecture documentation. That single initiative accounted for roughly a third of the day's output. The rest split across a full-stack task management application, product catalog metadata expansion, business planning document overhauls, and a resume parsing tool built from a design spec.

Task Log

# Task Human Est. Claude Supv. LF SLF
1 Task management backend core infrastructure: 35 files (models, schemas, config, auth, migrations, tests) 16h 6m 3m 160x 320x
2 Backend certification system: migration, model, schemas, service, and API endpoints 20h 8m 3m 150x 400x
3 Cross-platform user profile system: design and implementation across 6 repositories (backend services, web, desktop, iOS, architecture docs) with certification verification and enrollment flow 120h 55m 10m 131x 720x
4 Full-stack task management application: 150 files, backend and frontend 40h 25m 5m 96x 480x
5 Comprehensive business planning update: 10 documents rewritten for revised pricing model and enterprise launch roadmap 40h 25m 5m 96x 480x
6 Resume parsing tool from design spec: 81 files, 7,351 LOC, 6 parsers, 4 renderers, website generator, LLM integration, 129 tests 80h 50m 5m 96x 960x
7 Backend user profile system Phase A: migration and 13 files (models, schemas, services, endpoints for profile, avatar, privacy, account) 28h 18m 5m 93x 336x
8 Web application user profile page: profile view, avatar upload, certification card, API client, app integration 28h 18m 5m 93x 336x
9 Shared product info component with 3-client implementation (shared library, web refactor, desktop, iOS) 16h 12m 5m 80x 192x
10 Task management frontend scaffold: 57 files (React 19, Vite 6, TypeScript) with components, pages, hooks, API layer, auth, CSS modules 16h 12m 3m 80x 320x
11 Task management backend API routes and services: 19 files (7 services, 10 routes, exceptions, main configuration) 16h 12m 5m 80x 192x
12 Recruitment platform Phase 2: auth middleware, recruiter profile, layout shell, dashboard, settings (18 files) 20h 18m 2m 67x 600x
13 Desktop application user profile screen: types, auth client, IPC, preload, renderer API, profile component, CSS, app integration 20h 18m 5m 67x 240x
14 iOS task management repo creation with requirements and design documents derived from web counterparts 8h 8m 3m 60x 160x
15 Product catalog metadata expansion: exam info fields added to 40 domain specification files 16h 18m 3m 53x 320x
16 iOS user profile view: profile models, auth client actor, profile view, certification card, add-certification view, app state extensions, navigation wiring 20h 25m 5m 48x 240x
17 Recruitment platform Phase 1: project scaffolding, database schema, app shells (57 files, 12 tables) 16h 22m 3m 44x 320x
18 Business document updates: acquisition scenarios, enterprise training docs with revised pricing and valuations 8h 12m 5m 40x 96x
19 Web application enrollment flow wiring for product info landing page 4h 6m 3m 40x 80x
20 Desktop product info screen: screen component, CSS, i18n for 4 locales, app and dashboard wiring 3h 5m 3m 36x 60x
21 Valuation and funding document rewrite for pre-launch reality with 9 major revisions 3h 5m 3m 36x 60x
22 Backend billing endpoints: portal, plan changes, invoice retrieval 3h 5m 3m 36x 60x
23 Full-stack product info page: engine models, gateway, types, React component, CSS, i18n across 4 locales 24h 45m 8m 32x 180x
24 iOS product info view with new API model structs and optional type fixes 3h 6m 5m 30x 36x
25 SQLite import feature for task management: service, route, schema, main configuration 4h 8m 3m 30x 80x
26 Pitch deck rewrite: 3 files updated for bootstrapped growth plan, first raise, and Series B 4h 8m 5m 30x 48x
27 User enrollment tracking system: migration, model, schema, service, API, registration flow 2h 4m 3m 30x 40x
28 Auth bypass mode for internal tracking tool (backend, frontend, Docker configuration) 1.5h 4m 3m 22x 30x
29 Unit tests for SQLite import service: 31 tests covering helpers, database reads, and full async import 2h 6m 3m 20x 40x
30 iOS enrollment flow wiring for product info landing page (6 files) 4h 12m 3m 20x 80x
31 Desktop enrollment flow wiring for product info landing page 4h 12m 3m 20x 80x
32 Manifest patching pipeline: 3 rounds of patches for metadata and taxonomy fields across 40+ domain files 8h 25m 5m 19x 96x
33 Trademark symbol replacement (™ to ®) across 3 client applications (32 files) 2h 8m 3m 15x 40x
34 Auth bypass mode for leverage tracking tool (backend, frontend, Docker Compose) 1h 4m 3m 15x 20x
35 Import UI component for task management frontend: button and modal 2h 8m 3m 15x 40x

Legend: Human Est. = estimated human-equivalent time. Claude = wall-clock minutes for Claude to complete. Supv. = minutes I spent writing the prompt. LF = leverage factor (human time / Claude time). SLF = supervisory leverage factor (human time / my time).

Aggregate Statistics

Metric Value
Total tasks 35
Total human-equivalent hours 602.5
Total Claude minutes 533 (8.9 hours)
Total supervisory minutes 142 (2.4 hours)
Total tokens consumed ~3,603,000
Weighted average leverage factor 67.8x
Weighted average supervisory leverage factor 254.6x

Analysis

The cross-platform user profile system (task 3, 131x) was the largest single task of the day: 120 human-equivalent hours compressed into 55 minutes. The implementation spanned two backend services, three client applications, and architecture documentation. Profile pages, certification verification, enrollment flows, and GDPR-compliant data handling all had to stay consistent across web, desktop, and iOS. A team of three engineers (backend, frontend, iOS) would need two full weeks to deliver that scope with proper cross-platform parity. Ten minutes of prompt writing produced the same result.

The resume parsing tool (task 6, 96x) hit the highest supervisory leverage of the day at 960x. Five minutes of prompting yielded 81 files, 7,351 lines of code, 6 document parsers, 4 output renderers, a website generator, LLM integration, and 129 passing tests. That ratio reflects the nature of greenfield builds from a clear design spec: Claude can hold the full architecture in context and generate internally consistent code across dozens of files without the constant context-switching that slows human developers.

The task management application followed a deliberate three-phase pattern. Backend infrastructure first (task 1, 160x): 35 files including models, schemas, configuration, auth, and migrations in 6 minutes. Frontend scaffold next (task 10, 80x): 57 React 19 files in 12 minutes. Then the full application assembly (task 4, 96x): 150 files in 25 minutes. Breaking the build into phases like this consistently produces cleaner architecture than asking for everything at once, and the total Claude time (43 minutes) is still a fraction of the 72 human-equivalent hours.

Business planning and investor documentation (tasks 5, 18, 21, 26) ran 30-96x. These documents require understanding product strategy, market positioning, and financial projections. Claude handles the structured writing well; the supervisory cost is in reviewing the output for strategic accuracy rather than writing quality.

The floor was the auth bypass mode for the leverage tracker (task 34, 15x) and the import UI component (task 35, 15x). Small, focused changes that require more reading and understanding of existing code than generating new code. Even at 15x, four minutes beats an hour.

Let's Build Something!

I help teams ship cloud infrastructure that actually works at scale. Whether you're modernizing a legacy platform, designing a multi-region architecture from scratch, or figuring out how AI fits into your engineering workflow, I've seen your problem before. Let me help.

Currently taking on select consulting engagements through Vantalect.