[{"data":1,"prerenderedAt":293},["ShallowReactive",2],{"project-directory-en":3},[4,61,108,149,193,229,272],{"id":5,"title":6,"body":7,"category":40,"description":34,"extension":41,"featured":42,"maturity":43,"meta":44,"navigation":42,"order":45,"path":46,"period":47,"repositoryUrl":48,"seo":49,"slug":50,"status":51,"stem":50,"summary":52,"technologies":53,"__hash__":60},"projects_en\u002Flucidify.md","Lucidify",{"type":8,"value":9,"toc":33},"minimark",[10,15,19,23,26,30],[11,12,14],"h2",{"id":13},"product-focus","Product focus",[16,17,18],"p",{},"Lucidify explores a private journaling workflow for recording and revisiting dreams. The repository contains a Vue interface, account and dream API routes, and MongoDB models for users and dream entries.",[11,20,22],{"id":21},"implementation-and-decisions","Implementation and decisions",[16,24,25],{},"I worked on a client\u002Fserver application rather than a static journal: authenticated routes scope dream operations to a user, while a separate AI service handles optional text analysis through the backend. Keeping that integration behind an API route means the feature depends on server configuration; the route reports that it is unavailable when its API key is missing.",[11,27,29],{"id":28},"status-and-limitations","Status and limitations",[16,31,32],{},"This is a prototype codebase, not evidence of a validated or production service. Dream analysis is generated text and should not be presented as scientific, psychological, or medical advice. I did not verify a complete deployed user flow or any adoption outcomes.",{"title":34,"searchDepth":35,"depth":35,"links":36},"",2,[37,38,39],{"id":13,"depth":35,"text":14},{"id":21,"depth":35,"text":22},{"id":28,"depth":35,"text":29},"selected project","md",true,"8",{},"1","\u002Flucidify","2023-2026","https:\u002F\u002Fgithub.com\u002FKarolGitHub\u002Flucidify\u002Ftree\u002Fbb429c2f963262e43a98126b903b2faa53acded5",{"title":6,"description":34},"lucidify","live","A Vue dream journal with authenticated accounts, dream records, and backend-routed text analysis.",[54,55,56,57,58,59],"Vue 3","TypeScript","Node.js","Express","MongoDB","OpenAI API","Yc4r3qiMdznSRdQ4G46ZS0-LoAt6FZEKK-i0G0KzKeE",{"id":62,"title":63,"body":64,"category":40,"description":34,"extension":41,"featured":42,"maturity":93,"meta":94,"navigation":42,"order":95,"path":96,"period":97,"repositoryUrl":98,"seo":99,"slug":100,"status":51,"stem":100,"summary":101,"technologies":102,"__hash__":107},"projects_en\u002Fgeomapping-studio.md","GeoMapping Studio",{"type":8,"value":65,"toc":88},[66,70,73,75,78,80],[11,67,69],{"id":68},"problem","Problem",[16,71,72],{},"GeoJSON is useful but difficult to inspect and edit as raw data. GeoMapping Studio puts mapping and feature editing in a visual workspace, with a table view for inspecting feature properties.",[11,74,22],{"id":21},[16,76,77],{},"The React application keeps map, drawing, selection, and view state at the app level, then passes it to a dedicated map view. That view combines DeckGL with editable GeoJSON layers from nebula.gl. Separate hooks and services load GeoJSON from a URL and export the current collection as a file, keeping these data operations apart from the map component.",[11,79,29],{"id":28},[16,81,82,83,87],{},"The checked-in code supports drawing and editing map features, URL-based GeoJSON loading, table inspection, and export. The map component explicitly disables TypeScript checking and uses ",[84,85,86],"code",{},"any"," for several map values, so I do not claim strict type safety. The public client-side demo shell did not verify the full mapping workflow; the pinned source revision is the evidence for this description.",{"title":34,"searchDepth":35,"depth":35,"links":89},[90,91,92],{"id":68,"depth":35,"text":69},{"id":21,"depth":35,"text":22},{"id":28,"depth":35,"text":29},"7",{},"2","\u002Fgeomapping-studio","2025","https:\u002F\u002Fgithub.com\u002FKarolGitHub\u002Fgeomapping-studio\u002Ftree\u002F49a312f98b75c934ba936725edf0f69a3e077533",{"title":63,"description":34},"geomapping-studio","An interactive mapping workspace for drawing, editing, importing, and exporting GeoJSON features.",[103,55,104,105,106],"React","deck.gl","nebula.gl","Mapbox GL","pIkgJwNwq0BMzXXjh9lVp5xVd0DUTB8ISSjIpihxVf8",{"id":109,"title":110,"body":111,"category":40,"description":34,"extension":41,"featured":133,"maturity":134,"meta":135,"navigation":42,"order":136,"path":137,"period":97,"repositoryUrl":138,"seo":139,"slug":140,"status":51,"stem":140,"summary":141,"technologies":142,"__hash__":148},"projects_en\u002Fcoalescent-ai.md","CoalescentAI",{"type":8,"value":112,"toc":128},[113,115,118,120,123,125],[11,114,69],{"id":68},[16,116,117],{},"CoalescentAI brings several team workflows—whiteboards, tickets, and customer records—into one application rather than treating collaboration as a standalone canvas.",[11,119,22],{"id":21},[16,121,122],{},"The repository separates a Next.js client from a Fastify service. The server exposes a tRPC router and uses Socket.IO for whiteboard events; when a client joins a board, the inspected handler loads its saved events from Supabase. On the client, the whiteboard, ticket, and customer screens are separate routes, keeping the workspace areas distinct.",[11,124,29],{"id":28},[16,126,127],{},"The codebase is a prototype, not a verified production collaboration service. The public deployment currently reaches a sign-in screen; I did not create an account or validate authentication, persistence, or multi-user synchronization against a running backend. The repository code is the evidence for the implementation described here.",{"title":34,"searchDepth":35,"depth":35,"links":129},[130,131,132],{"id":68,"depth":35,"text":69},{"id":21,"depth":35,"text":22},{"id":28,"depth":35,"text":29},false,"5",{},"3","\u002Fcoalescent-ai","https:\u002F\u002Fgithub.com\u002FKarolGitHub\u002Fcoalescent-ai\u002Ftree\u002F9102344dbc2afd740b7da7819ce610f0b39884cd",{"title":110,"description":34},"coalescent-ai","A collaboration workspace prototype with whiteboard, ticket, and customer interfaces backed by a TypeScript service.",[143,103,55,144,145,146,147],"Next.js","Fastify","tRPC","Socket.IO","Supabase","amoO2XJxmTD2sWdxiIo5gHYeopjHKo2dAO0EtRUaBCg",{"id":150,"title":151,"body":152,"category":40,"description":34,"extension":41,"featured":133,"maturity":43,"meta":178,"navigation":42,"order":179,"path":180,"period":181,"repositoryUrl":182,"seo":183,"slug":184,"status":185,"stem":184,"summary":186,"technologies":187,"__hash__":192},"projects_en\u002Fstate-management-benchmark.md","State Management Benchmark",{"type":8,"value":153,"toc":173},[154,158,161,165,168,170],[11,155,157],{"id":156},"research-question","Research question",[16,159,160],{},"This historical project explores how several framework implementations handle the same interactive table operations. The repository includes React, Svelte, Vue, and Angular apps with controls for creating, adding, loading, updating, swapping, selecting, and deleting generated rows.",[11,162,164],{"id":163},"measurement-approach","Measurement approach",[16,166,167],{},"The apps mark and measure UI operations with the browser Performance API. They also collect navigation, resource, paint, and (where available) memory timing information, and allow operation timings to be exported as CSV. This makes the repository useful as a record of an experiment, not a current framework ranking.",[11,169,29],{"id":28},[16,171,172],{},"Repository activity is from 2021–2022. The checked-in materials do not provide a reproducible cross-browser protocol, controlled environment, or published comparative results, and the implementations use different framework versions. I therefore make no performance winner or present-day framework claim from this work.",{"title":34,"searchDepth":35,"depth":35,"links":174},[175,176,177],{"id":156,"depth":35,"text":157},{"id":163,"depth":35,"text":164},{"id":28,"depth":35,"text":29},{},"4","\u002Fstate-management-benchmark","2021–2022","https:\u002F\u002Fgithub.com\u002FKarolGitHub\u002Fstate-management-benchmark\u002Ftree\u002Fea67883b2e4c9d6b45d910ab04647b4193a69ca8",{"title":151,"description":34},"state-management-benchmark","completed","A historical browser experiment comparing common data-table operations across React, Svelte, Vue, and Angular applications.",[103,188,189,190,191],"Svelte","Vue","Angular","Web Performance API","XGK7-3kR9Vg43zx7WBS3oLzQUraoIGsiox-vAeQxU88",{"id":194,"title":195,"body":196,"category":40,"description":34,"extension":41,"featured":133,"maturity":134,"meta":218,"navigation":42,"order":134,"path":219,"period":97,"repositoryUrl":220,"seo":221,"slug":222,"status":51,"stem":222,"summary":223,"technologies":224,"__hash__":228},"projects_en\u002Fproofmint-ai.md","ProofMintAI",{"type":8,"value":197,"toc":213},[198,200,203,205,208,210],[11,199,69],{"id":68},[16,201,202],{},"ProofMintAI explores a document workflow that combines file handling with storage and verification concepts. The inspected frontend includes a file-upload interface with public and encrypted storage options.",[11,204,22],{"id":21},[16,206,207],{},"The backend has IPFS upload and retrieval routes backed by a storage service. That service contains AES-256-GCM encryption and decryption helpers and Web3.Storage upload code, separating file storage operations from the Vue\u002FQuasar interface.",[11,209,29],{"id":28},[16,211,212],{},"This is a prototype, and I did not verify an end-to-end upload, decryption, or external storage session. In particular, the document-hash route currently returns a placeholder response. I therefore do not describe the repository as providing verified notarization, a working blockchain proof, or production-ready document security.",{"title":34,"searchDepth":35,"depth":35,"links":214},[215,216,217],{"id":68,"depth":35,"text":69},{"id":21,"depth":35,"text":22},{"id":28,"depth":35,"text":29},{},"\u002Fproofmint-ai","https:\u002F\u002Fgithub.com\u002FKarolGitHub\u002Fproofmint-ai\u002Ftree\u002F83be42fc817ffce5741c897d06032482c0dfeae3",{"title":195,"description":34},"proofmint-ai","A document workflow prototype with public and encrypted IPFS upload and retrieval code paths.",[189,225,56,57,226,227],"Quasar","IPFS","AES-256-GCM","JSUrxUzKPEn31KqHcRftXDrHcNPxPgrUUE6TPqELARw",{"id":230,"title":231,"body":232,"category":40,"description":34,"extension":41,"featured":133,"maturity":134,"meta":258,"navigation":42,"order":259,"path":260,"period":261,"repositoryUrl":262,"seo":263,"slug":264,"status":51,"stem":264,"summary":265,"technologies":266,"__hash__":271},"projects_en\u002Fcrmaize.md","CRMAIze",{"type":8,"value":233,"toc":253},[234,236,239,241,248,250],[11,235,69],{"id":68},[16,237,238],{},"CRMAIze explores customer-retention workflows for e-commerce teams: grouping customer records and using those groups in campaign-management screens.",[11,240,22],{"id":21},[16,242,243,244,247],{},"The application is organized around PHP controllers, repositories, and services, with Twig-rendered pages. Its ",[84,245,246],{},"AIService"," groups customers using explicit purchase and recency thresholds, and the campaign controller requires authentication before rendering campaign workflows. This makes the current segmentation logic understandable and rule-based rather than dependent on an opaque model.",[11,249,29],{"id":28},[16,251,252],{},"The inspected segmentation service uses fixed business rules; I do not claim trained machine learning or validated churn predictions. When checked, the configured public deployment returned a database connection error, so I am not linking it as a working demo. The source revision is available for review.",{"title":34,"searchDepth":35,"depth":35,"links":254},[255,256,257],{"id":68,"depth":35,"text":69},{"id":21,"depth":35,"text":22},{"id":28,"depth":35,"text":29},{},"6","\u002Fcrmaize","2025-2026","https:\u002F\u002Fgithub.com\u002FKarolGitHub\u002FCRMAIze\u002Ftree\u002F075efd63faaf736c582a014864ddf609f048a528",{"title":231,"description":34},"crmaize","A PHP\u002FTwig customer-retention application with customer segmentation and campaign-management workflows.",[267,268,269,270],"PHP","Twig","PostgreSQL","JavaScript","T9lBznRar3CyH2VtGAjh7gSPGcz7Fo7yznkT1iGlsxM",{"id":273,"title":274,"body":275,"category":282,"description":279,"extension":41,"featured":42,"maturity":95,"meta":283,"navigation":42,"order":93,"path":284,"period":285,"repositoryUrl":286,"seo":287,"slug":288,"status":289,"stem":288,"summary":290,"technologies":291,"__hash__":292},"projects_en\u002Frender-tally.md","RenderTally",{"type":8,"value":276,"toc":280},[277],[16,278,279],{},"RenderTally is a separate planned application, not the name of this company. Its intended initial focus is supporting one-time performance engagements. It is not live, and this website does not provide a scanner, client portal, dashboard, or subscription service.",{"title":34,"searchDepth":35,"depth":35,"links":281},[],"personal project",{},"\u002Frender-tally","2026-2027",null,{"title":274,"description":279},"render-tally","in development","A planned private workbench for expert-led website performance audits and improvement engagements.",[],"2cyieW3YSp-tN9epu9n1FTitssGs0F1T0KOs58j0xrM",1792326467203]