FOKRUL ISLAM
AI Engineer·Multimodal AI & Agents
Human-centric AI systems, local-first agents, and future intelligent infrastructure.
Building AI systems that run privately, reason deeply, and remain under full human control — no cloud lock-in, no black boxes.
Architecting the Synergy of Anthropology & AI
Powerful systems should feel simple. The goal is AI that hides its complexity behind calm, human-friendly interfaces — built for real people with real needs.
Local-First Intelligence
AI systems that run privately on your machine — no cloud dependency, no data leaving your control.
Multimodal Agents
Agents that see, read, and reason across text, images, and documents simultaneously.
Human-Centric Design
Complex systems hidden behind calm, simple interfaces. Powerful inside. Quiet outside.
High-Performance Stacks
Rust for speed and safety. Python for intelligence. Together they form the backbone of real AI products.
"Today we build AI agents and tools. Tomorrow the same intelligence layer expands into physical systems — robotics, drones, satellite awareness. One coherent intelligence. Everywhere."
— FOKRUL ISLAM, Founder of Intelliyah AI
Built in Public. Real Systems.
Not prototypes. These are engineered systems with real AI stacks, real distributed architectures, and real use cases.
Aivora
Autonomous Multimodal AI OS with Supabase pgvector RAG, local embeddings, citations, and WebLLM-ready browser AI.
Intelliyash
Self-optimizing local AI runtime for low-end machines, designed to detect hardware and select the best model automatically.
Clairvoyant
Multimodal Vision AI agent bridging Anthropology and Tech to assist visually impaired users.
distributed-ai-tensor-runtime
Rust-based distributed tensor runtime core with Candle compute, P2P foundation, ResourceGuard, WASM policy, telemetry, and dashboard.
Synapse Overlord
Local-first offline AI project builder dashboard using Rust, Axum, Ratatui, and deterministic local generation.
Sapience-AI
Local-first AI wellness monitoring system using Flutter, FastAPI, and Python local AI.
Intelliyah AI Products
A growing suite of tools built on a single principle: AI should be accessible, private, and genuinely useful.
Intelliyah AI
Universal AI agent builder for everyone.
Build, configure, and deploy intelligent AI agents without writing a single line of backend code.
Idea-to-Local-AI Builder
Drop an idea and generate a local-first AI assistant.
Describe your goal. Intelliyah designs the assistant, prepares the knowledge workflow, and packages it for local deployment.
AI Knowledge Vault
Store documents, memory, and vector knowledge for custom AI agents.
Securely store documents, notes, agent memory, chat history, and vector-based knowledge for AI retrieval.
Intelliyash Runtime
Future local AI runtime engine for low-end and private machines.
Detects hardware capabilities and selects the optimal model. Designed for devices that cannot afford cloud dependency.
Bring Your Idea. I'll Build the AI.
From simple AI assistants to local-first agent platforms and full-stack AI SaaS products — I help turn ideas, documents, and workflows into intelligent systems.
Custom AI Agent Development
End-to-end intelligent agent systems tailored to your workflow.
PDF / Document Chatbot
Conversational AI that reads, understands, and answers from your documents.
RAG Knowledge Base AI
Vector-powered retrieval systems that give AI accurate, grounded memory.
Local-First AI Workflow
Private, offline-capable AI systems — no cloud dependency.
Full-Stack AI Web App
React/Next.js + Python/Rust AI apps from design to deployment.
AI SaaS MVP Development
Launch your AI product idea as a polished SaaS in weeks.
Multimodal AI Prototype
Vision, audio, and text AI systems working together in one product.
AI Knowledge Storage / Memory
Persistent agent memory systems with vector search and retrieval.
Rust/Python AI Engine Architecture
High-performance AI infrastructure built with Rust and Python.
Custom AI Agent Training
Knowledge training for custom agents using RAG, vector memory, documents, and workflow rules.
AI Workflow Automation
Connect AI agents with n8n, Supabase Functions, Python workers, and business pipelines.
n8n Workflow Integration
Automated workflows connecting apps, APIs, databases, and AI agents via n8n.
Business Process Automation
End-to-end automation of business workflows with AI classification, routing, and actions.
RAG + Vector Memory Setup
Complete setup of retrieval-augmented generation with Supabase pgvector and embeddings.
Train Agents With Your Knowledge.
Not every business needs a new AI model from scratch. Most need a custom AI agent that understands their documents, products, rules, and workflows.
Intelliyah is designed to turn business knowledge into agent memory using RAG, vector search, prompt rules, and tool-connected workflows — so agents answer accurately from your data, not from generic training data.
What agent training actually means
Document Knowledge Training
PDFs, notes, policies, research files, product catalogs, FAQs, and business documents — uploaded and indexed into agent memory.
RAG + Vector Memory
Embeddings, Supabase pgvector, and semantic search power accurate knowledge retrieval — agents answer from your data, not generic training.
Workflow Rules & Prompts
Teach agents how to follow business logic, response style, task sequences, escalation rules, and domain-specific decision boundaries.
Tool-Connected Agents
Prepare agents to work with email, forms, databases, dashboards, APIs, and automation tools — moving beyond chat into real workflows.
Advanced Fine-Tuning
Future / CustomFuture and custom enterprise option: train or fine-tune language models on your dataset when the GPU capacity, evaluation pipeline, and data volume are ready. This is an advanced service, not a standard offering.
Ready to train an agent on your knowledge?
Share your documents, workflow, or use case. I'll design a custom agent training pipeline that fits your business.
Start Agent TrainingConnect AI Agents With Real Business Tools.
Custom AI becomes more powerful when it can act inside a workflow — not just chat.
Intelliyah can connect agents with forms, email, databases, dashboards, notifications, and business pipelines using n8n, Supabase Functions, custom Python workers, or future Rust automation services.
Automation examples
Automation tools & integrations
n8n Automation
Open workflow automation for connecting apps, APIs, databases, and AI agents. Self-hostable, no vendor lock-in.
Supabase Edge Functions
Database-triggered backend automation, cron jobs, secure server logic, and real-time database events.
Python Workers
AI processing, document parsing, embeddings, summaries, and background tasks — flexible and production-ready.
Rust Automation Engine
Future local-first, high-performance automation for package building, system orchestration, and low-latency workflows.
Zapier / Make / Pipedream
Alternative automation platforms for client-specific workflows, third-party integrations, and no-code pipelines.
Node-RED
Future-friendly option for IoT, robotics, sensors, and local hardware automation integrated with AI agents.
Automate your AI-powered workflow.
Bring your workflow idea. I'll design the automation pipeline and connect your AI agents to the right tools for your business.
From Idea to Local AI in Minutes.
Drop your idea. Intelliyah understands the goal, designs a custom AI assistant, prepares the knowledge workflow, and previews a local-first package structure.
Drop Your Idea
Describe your goal in plain language. What do you want your AI to do?
Intelliyah Designs It
The engine understands your goal, designs the assistant, and prepares the knowledge workflow.
Package Generated
A complete local-first AI package is assembled — ready to install and run privately.
Run Locally
Install the package. Run your custom AI assistant. No API key. No cloud. No lock-in.
Your Data. Your AI Memory.
Securely store documents, notes, agent memory, chat history, and vector-based knowledge so custom AI agents can search, remember, and answer from user-owned data.
Document Storage
Upload PDFs, notes, and files. Your AI will read and remember them.
Agent Memory
Persistent memory so your AI learns from every conversation.
Vector Search
Semantic search across your knowledge base — AI finds what matters.
Private by Design
Your data stays in your vault. No external transmission required.
Local-First Storage
Runs on your own infrastructure. Cloud sync is optional.
RAG Retrieval
Accurate, grounded answers pulled from your own documents.
Drop documents, PDFs, or notes here
Supported: PDF, TXT, MD, DOCX
product-requirements.pdf
2.4 MB
meeting-notes-2025.md
48 KB
knowledge-base.txt
1.1 MB
3
Documents
1,284
Vectors
47
Searches
Next Vision: Self-Boot Virtual Self
Beyond custom AI agents, the next vision is a Self-Boot Virtual Self — a personal AI layer designed to understand a person’s knowledge, workflow, writing style, and goals. It begins as a digital shadow agent for task assistance, document reasoning, memory, and workflow support, with long-term research toward physical robotic embodiment.
Digital-first vision now. Physical embodiment remains a long-term research direction.
Personal Knowledge Memory
Learns from documents, notes, workflows, and project context — building a persistent intelligence layer that understands your domain at a deep level.
Shadow Agent Workflow
A parallel agent layer that assists with tasks, planning, writing, research, and decision support — running alongside your work without disrupting it.
Local-First Personal AI
Private memory, local intelligence, and user-owned data — your digital self stays on your hardware, fully portable, zero cloud exposure.
Future Robotic Embodiment
Long-term research direction for extending digital intelligence into physical assistants — bridging software agents and real-world automation. Not a current product.
Intelligence That Expands Everywhere.
Today, Intelliyah starts with local-first AI agents, knowledge systems, and human-centric software. Tomorrow, the same intelligence layer can expand into physical and spatial systems — AI-assisted robotics, autonomous drone workflows, and satellite-aware intelligence systems.
AI-Based Robotics
Exploring intelligent physical assistants and automation systems that leverage the same local-first AI intelligence layer built for software agents.
The same reasoning and perception systems powering Intelliyah agents could, in future research, be adapted for robotic perception, planning, and action execution.
AI-Assisted Drones
Investigating AI-native drone systems capable of autonomous perception, terrain navigation, inspection workflows, and field intelligence.
Future research direction for applying multimodal vision models and edge AI inference to unmanned aerial systems for inspection, monitoring, and field intelligence.
Satellite-Aware Intelligence
Long-term research into geospatial AI, remote sensing data analysis, and large-scale environmental intelligence systems.
Exploring how AI can process satellite imagery and geospatial data for environmental monitoring, infrastructure analysis, and planetary-scale intelligence.
Software intelligence today. Physical intelligence tomorrow.
Robotics is a long-term direction for extending AI agents into real-world assistants and automation systems.
From local AI agents to future intelligent systems — the journey starts here.
Every tool built today is a foundation for the intelligent systems of tomorrow. Intelliyah is the launchpad.
The Stack Behind Real AI Systems.
Every tool chosen for a reason. Performance, privacy, and human-centricity guide every architectural decision.
AI & Intelligence
Systems & Performance
Backend & APIs
Frontend & UI
Data & Memory
DevOps & Tools
Stack Philosophy
Rust for correctness and performance where it matters. Python for rapid AI prototyping and ML ecosystem access. TypeScript for safe, scalable frontends. The combination produces systems that are both fast and friendly.
Let's Build Something Intelligent Together.
Have an AI project in mind? Need a custom agent, document chatbot, or full AI SaaS? Reach out — let's explore what's possible.
fokrulanthro16@gmail.com
Best for project inquiries and detailed proposals
+8801732457882
For quick questions and project discussions
fokrulanthro16-eng
Browse real engineering projects and code
@IslamLectu91999
AI engineering updates and thoughts
Not sure where to start?
Tell me your idea in one sentence. I'll tell you exactly how I'd build it — and what tech stack would make it fast, private, and maintainable.
Send Your Idea