Yossi Elkrief MaTriXy

Inferal Workspace Architecture

Your org's brain that AI can use

This document describes the conceptual architecture of the Inferal Workspace - a text-based, version-controlled knowledge and operations hub designed to replace tools like Notion and Webflow while being natively accessible to AI assistants.

Why We Built This

As an engineering-driven organization, we found ourselves fighting our tools instead of using them. Notion couldn't keep up with how we actually work - context scattered across pages, no version control, and AI that could read but not act. Webflow meant our website lived in a silo, disconnected from our codebase and deployment pipelines. Every tool was another tab, another context switch, another place where knowledge went to die.

Agent Learning via Early Experience + ACE Integration

Production-Ready Framework for Continuous Agent Learning

A complete implementation of reward-free reinforcement learning through world modeling, exploration, and self-reflection, with full ACE (Adaptive Context Engineering) integration for knowledge curation and semantic deduplication.

What Is This?

<core_identity> You are an assistant called Cluely, developed and created by Cluely, whose sole purpose is to analyze and solve problems asked by the user or shown on the screen. Your responses must be specific, accurate, and actionable. </core_identity>

<general_guidelines>

NEVER use meta-phrases (e.g., "let me help you", "I can see that").
NEVER summarize unless explicitly requested.
NEVER provide unsolicited advice.
NEVER refer to "screenshot" or "image" - refer to it as "the screen" if needed.
ALWAYS be specific, detailed, and accurate.

	//
	// ContentView.swift
	// Airdrop Demo
	//
	// Created by Daniel Kuntz on 7/30/23.
	//

	import SwiftUI

	struct ContentView: View {

	# API Stress Test

	Create a Bash-based stress testing suite that simulates real user interactions with
	your API and measures system performance under load.

	## Core Script (`api_stress_test.sh`)
	- Simulate realistic user flows with sequential API calls (GET, POST, PUT, DELETE)
	- DELETE requests should be executed last to clean up test data
	- Support bearer token authentication and custom headers
	- Base URL should be configurable

	#ifndef MarchingCubesParams_h
	#define MarchingCubesParams_h
	#include <simd/simd.h>

	struct MarchingCubesParams {
	simd_uint3 cells;
	simd_float3 origin;
	simd_float3 cellSize;
	float isoLevel;
	simd_float3 centerA;

	# export OPENAI_API_KEY=sk_proj-...
	# uv run gpt-5-voice-agent.py

	# /// script
	# dependencies = [
	# "numba==0.61.2",
	# "openai==1.99.1",
	# "python-dotenv",
	# "fastapi[all]",
	# "uvicorn",

	Hi LinkedIn friend!

	Here is how I managed to have the agent I'm building reduce token usage.

	1. My agent has 62 tools (and growing quickly in terms of number of tools..)
	2. Each tool has a description. All in all I was sending the entire 62 tools+description in every agent turn.

	It came out to 10k tokens before even the system prompt+user prompt - ON EVERY TURN.

	The solution I found was to do a preflight LLM request to select only the relevant tools for the user request.

	import asyncio
	import os
	import sys
	from dataclasses import dataclass

	import aiohttp
	import google.ai.generativelanguage as glm
	from dotenv import load_dotenv
	from loguru import logger
	from runner import configure

	# train_grpo.py
	import re
	import torch
	from datasets import load_dataset, Dataset
	from transformers import AutoTokenizer, AutoModelForCausalLM
	from peft import LoraConfig
	from trl import GRPOConfig, GRPOTrainer

	# Load and prep dataset