You will get a private AI that reads your documents, audio and video, on your server
Rising Talent
Rising Talent
Project details
You will get a private AI system that reads your documents, listens to your audio recordings and indexes your video - running on your own server, or on a dedicated instance I set up for you. Nothing is sent to OpenAI, Google or any other external AI provider: the model runs on infrastructure you control, in the country or region you choose. Every answer cites its exact source - file name, page or timestamp - so you can always verify it. Audio and video are transcribed locally and become searchable word by word. I build and ship production systems and automation pipelines, and I work asynchronously in written English, French, Spanish and Italian. When the project ends, all data is deleted and the instance is wiped.
AI Algorithms
Large Language Model, Multimodal Large Language Model, Transformer ModelAI Applications
AI Chatbot, AI-Enhanced Classification, Automatic Speech Recognition, Natural Language UnderstandingAI Development Language
PythonAI Models
LLaMA, WhisperWhat's included
| Service Tiers |
Starter
$390
|
Standard
$890
|
Advanced
$1,900
|
|---|---|---|---|
| Delivery Time | 5 days | 10 days | 21 days |
Number of Revisions | 1 | 2 | 3 |
AI Model Integration | - | - | - |
Batch Normalization | |||
Database Integration | |||
Detailed Code Comments | |||
Image Upscaling | - | - | - |
MLOps | - | - | - |
Model Deployment | - | - | - |
Model Documentation | - | - | - |
Model Monitoring | - | - | - |
Model Testing & Optimization | - | - | - |
Model Tuning | - | - | - |
Natural Language Processing | - | - | - |
NLP Tokenization | - | - | - |
Pre-Training | - | - | - |
Prompt Engineering | - | - | - |
Setup File | - | - | - |
Source Code | - | - | - |
About Pasqualino
Email & Workflow Automation Systems | Next.js Full-Stack Developer
Rome, Italy - 10:20 pm local time
Most of my work is backend automation - workers that send, read, classify and update data 24/7 - plus full web applications built from scratch and kept alive in production. Not prototypes: everything below is online right now, and I can show you each one live.
EMAIL AND WORKFLOW AUTOMATION
A production daemon sending ~288 certified emails per day, with randomized 280-340s intervals, template rotation and automatic bounce handling. A companion worker reads the mailbox and classifies every reply - delivery receipts, NDR and bounces, human answers, out-of-office - then updates database records on its own, including auto-learning the sender's real address so future matching still works. SMTP, IMAP, STARTTLS, deliverability, bounce handling.
FULL-STACK WEB APPLICATIONS
Next.js (App Router), TypeScript, React, Prisma, MySQL, Tailwind. Deployed on my own Hetzner servers over SSH with pm2 and systemd. In production:
- A portal built on 174,000 Italian non-profit organizations imported from the public national registry: import, normalization, search over a real dataset.
- An internal CRM: records, send logs, replies, states, dashboard.
- A public association website with members area, course catalog with recurring slots, document downloads and structured data.
PROGRAMMATIC MULTILINGUAL SEO
I take a large dataset and turn it into a site that generates itself in dozens of languages and gets indexed by Google. Live examples I own and run:
- An anime and manga database, 250,000+ entries, 28+ languages
- A Twitch creator ranking site with user voting, Twitch OAuth login and Turnstile anti-fraud, 35 languages
- Video game, plant and mineral databases (one wired to Wikidata), plus a public administration guide site
The hard part is not generating pages - it is not generating garbage. The pipeline ships with anti-repetition checks (n-gram and opening-line comparison), word frequency validation and thin-content detection.
CONTENT AUTOMATION, END TO END
A full chain from news source to published post: news watcher, post generation with custom tone rules, carousel generation exported as paginated PDF, self-balancing queue and calendar, publishing through the official LinkedIn API (document upload included) and Telegram. Real numbers: 250 articles scheduled over ~2 months, one per category per weekday at randomized hours. One site alone holds ~1,017 published articles.
Honest scope: LinkedIn, Telegram and the website are fully connected. Instagram, TikTok, Facebook and X are not.
AI INSIDE PRODUCTS
LLM-based scoring of public funding calls. A public chatbot with a custom system prompt that answers differently to already-registered organizations. Real audio transcription with faster-whisper feeding an automated YouTube SEO package. 101 generated 4:5 posters shipped to production. I put models inside software that then runs on its own - not "I use ChatGPT".
GAMES
A live browser football MMO running on PM2. Three playable characters per account (player, manager, chairman), a shared multiplayer world, plus a single-player NPC mode whose world is stored on the user's device instead of the server.
BACKGROUND
19 published books, including 14 essays on artificial intelligence. 20 years teaching adults, currently top 10 in Italy on a major tutoring platform. Native Italian speaker.
HOW I WORK
Written English, fully async - no daily calls. Fixed-scope projects. I am direct about what I can and cannot do: if something in your brief is a bad idea, you hear it before we start, not after.
Tell me what should run without you, and I will tell you whether I can build it.ive browser football MMO running on PM2. Three playable characters per account (player, manager, chairman), a shared multiplayer world, plus a single-player NPC mode whose world is stored on the user's device instead of the server.
BACKGROUND
19 published books, including 14 essays on artificial intelligence. 20 years teaching adults, currently top 10 in Italy on a major tutoring platform. Native Italian speaker.
HOW I WORK
Written English, fully async - no daily calls. Fixed-scope projects. I am direct about what I can and cannot do: if something in your brief is a bad idea, you hear it before we start, not after.
Tell me what should run without you, and I will tell you whether I can build it.
Steps for completing your project
After purchasing the project, send requirements so Pasqualino can start the project.
Delivery time starts when Pasqualino receives requirements from you.
Pasqualino works on your project following the steps below.
Revisions may occur after the delivery date.
Data audit and setup
I look at a sample of your material, agree where the system will run and in which region, then provision the private instance and lock it down.
Ingestion and indexing
Your documents are loaded, audio and video are transcribed locally, and everything is indexed so the AI can search it and cite the exact source.