Early-Stage AI Startup

AI infrastructure, automation and intelligent developer tools.

Alsaidy AI Labs is an early-stage, self-funded AI startup building open-source and product-focused tools for AI model orchestration, developer automation, benchmarking, and specialized AI applications.

AI
Routing & Orchestration Pipeline
Routing Simulation
01. Ingestion
User Request
"Classify fungal morphology & synthesize pipeline"
Task: Bio AI Multi-StepPriority: Normal
CORE ROUTER
Alsaidy Model Router
Task Complexity:Reasoning Heavy
Routing Policy:Cost & Latency Aware
AWS Bedrock
Claude 3.5 Sonnet
SIMULATED
Anthropic Direct
Claude 3.5 Haiku
OpenAI
GPT-4o
Illustrative Routing Output:AWS Bedrock (Claude 3.5 Sonnet)
Policy: Cost-Aware·Tier: Balanced·Fallback: Enabled

Core Focus

What We Build

We're building modular software for AI model management, developer automation, and computational research.

AI Model Orchestration

Building a unified API layer to connect and manage multiple AI providers — AWS Bedrock, Anthropic, OpenAI, and self-hosted models — behind a single gateway.

Multi-Model Routing

Developing intelligent request routing that selects models based on task complexity and cost constraints, with automatic fallback between providers.

Agent & CLI Delegation

Creating task decomposition tools that allow agents to spawn, coordinate, and delegate subtasks across terminal and cloud environments.

Cost & Usage Analytics

Designing dashboards and logging infrastructure for per-request token accounting, latency tracking, and monthly spend projections.

AI Benchmarking

Building evaluation harnesses to measure latency, throughput, accuracy, and cost across different foundation models and providers.

Research Tools for Life Sciences

Applying AI and computational methods to microbiology — including fungal morphology analysis, genomic sequence processing, and lab data automation.

Flagship Project

AI Model Orchestrator

A platform for managing multiple AI providers and models through a single API — with routing, fallback, cost tracking, and benchmarking capabilities.

What We're Building

A single API gateway that abstracts across AWS Bedrock, Anthropic, OpenAI, and local models.

Automatic provider fallback — if one provider hits rate limits, traffic reroutes to the next available.

Intelligent routing that matches task complexity to the right model tier to reduce costs.

Per-request token usage logging and cost tracking across all connected providers.

Stack

PythonFastAPITypeScriptNext.jsAWS BedrockSupabase

Status: In Development — actively being built and iterated on.

Follow Development on GitHub
orchestrator.console.alsaidy-ai.online
Interface Preview — Simulated Provider State
Provider Connections
4 Planned
Bedrock · Claude · GPT · Local
Routing Mode
Dynamic Tier
Cost & Complexity Aware
Fallback Strategy
Automatic
Rate limit failover
Telemetry Tracking
Per-Query
Token · Latency · Cost
Provider Engine & Health MatrixAuto-Failover Planned
AWS Bedrock (Claude 3.5 Sonnet)
PRIMARY (TARGET)
Anthropic Claude 3.5 Haiku
STANDBY (FAILOVER 1)
OpenAI GPT-4o
STANDBY (FAILOVER 2)
Local Models (Ollama/vLLM)
OPTIONAL LOCAL

Open-Source & Active Work

Projects

Tools and frameworks currently being developed. Source code repositories are published on GitHub as each project reaches initial release.

Provider MeshPlanned Routing
AWS Bedrock
Anthropic
OpenAI
Failover: DynamicAPI Abstraction
In Development

AI Model Orchestrator

A platform for managing multiple AI providers and models through a single API — with routing, fallback, cost tracking, and benchmarking capabilities.

PythonFastAPITypeScriptNext.jsAWS BedrockSupabase
Development Repository (In Progress)
Semantic BranchingCost Optimized
Light Tasks→ Fast Tier (Haiku)
Complex Reasoning→ Sonnet / GPT-4o
In Development

Multi-Model Router

A routing proxy for LLM requests that selects the best provider based on cost and task complexity, with automatic fallback handling.

PythonFastAPIOpenAIAnthropicAWS Bedrock
Development Repository (In Progress)
agent-coordinator.sh
$ delegate --task "bio_eval"
├── worker_1 [genomics] [ASSIGNED]
└── worker_2 [eval] [PLANNED]
Early Stage

Agent Delegation CLI

A lightweight command-line toolkit for coordinating background agent tasks and managing subagent workflows.

TypeScriptNode.jsCLI
Development Repository (In Progress)
Eval DimensionsHarness Design
Latency MetricsTTFT & Total Duration
EconomicsCost-per-1k Output Tokens
Early Stage

AI Benchmarking Suite

An evaluation framework for measuring latency, throughput, cost, and output quality across different AI models and providers.

PythonDocker
Development Repository (In Progress)
Computer Vision PipelineResearch Prototype
Cyanobacteria / Spore Box
Morphology Segmentation
Research Prototype

Fungi.ai

A computer vision and genomic analysis pipeline focused on fungal morphology identification and microbial classification.

PythonPyTorchOpenCVBioPython
Development Repository (In Progress)

Developer Profile & Repositories

Active experiments, scripts, and work-in-progress on GitHub.

github.com/ahmed-makram-alsaidy

AI × Life Sciences

Research & Scientific Work

Bridging microbiology and software engineering — applying computational methods to biological problems and life sciences data.

AI × Environmental Microbiology

Applying computational methods and machine learning to microbial bioremediation, cyanobacteria growth dynamics, and pollutant degradation mechanisms.

AI Model Infrastructure

Designing multi-provider orchestration architectures, cost-optimization strategies, and autonomous agent evaluation frameworks.

Computational Bio-Pipelines

Building open-source tools for genomic sequence annotation, laboratory data processing, and microbiological image analysis.

Computational Microbiology & Cyanoremediation
Research Background — Faculty of Science, Damanhour University
AI × Environmental Microbiology
Bio-Vision Analyzer [Active Frame]Example Detection Overlay
Cyanobacteria Microscopic Filament [Structure A]
Bioremediation Target
Water Pollutant Adsorption Site [Metabolic Pathway]
Bio-Sorption Modeling
Scientific & Engineering Synthesis

Combining microbiology domain knowledge with computational and machine-learning pipelines to advance bioremediation research and microbial analysis.

Primary Focus: Microbial remediationPeer-Reviewed Backing

Published Research

Direct journal link
Cloud Architecture

Built with AWS

Our infrastructure architecture is being designed around Amazon Web Services for model access, distributed compute, telemetry, and research artifact storage.

AWS
Target Cloud Infrastructure Diagram
Proposed Architecture Blueprint
Client Layer
SDK & Web Clients

Python SDK, Next.js UI, & CLI Agent delegation tools.

HTTPS / JSON-RPC
Ingestion & Auth
Amazon API Gateway

SSL termination, request throttling, and token authentication.

Route Management
Compute & Routing
AWS Lambda / ECS

FastAPI router core, dynamic complexity classification, and failover logic.

Zero-Downtime Dispatch
Foundation Models
AWS Bedrock

Claude 3.5 Sonnet, Claude 3.5 Haiku, and specialized bio-reasoning models.

Primary Foundation Engine
Amazon S3 (Data & Artifact Store)

Stores benchmark evaluation runs, prompt regression test cases, fungal microscopy datasets, and genomic sequence cache.

Amazon CloudWatch (Telemetry & Audit)

Continuous latency logging, token spend monitoring, provider error alarms, and automated failover telemetry.

Foundation Model Layer

AWS Bedrock

Multi-model foundation model access and provider abstraction.

Serverless & Compute Layer

AWS Lambda / ECS

Planned compute layer for routing and background workloads.

Artifacts & Datasets

Amazon S3

Storage for evaluation artifacts and research datasets.

Observability & Metrics

Amazon CloudWatch

Observability, latency monitoring, and usage telemetry.

Infrastructure Roadmap & Resource Allocation

AWS Activate credits will support the development and testing of the platform's model orchestration, evaluation, and cloud infrastructure.

Stack

Technology

The providers, languages, and infrastructure we work with.

AI Providers & Planned Integrations

AWS BedrockAnthropicOpenAI

Languages & Frameworks

PythonTypeScriptNext.jsPyTorchFastAPITailwind CSS

Infrastructure

SupabaseVercelDocker

Leadership

Founder

AM

Ahmed Makram Siddek Alsaidy

Founder & Software Engineer

Microbiology researcher working at the intersection of AI, software engineering, and life sciences. Building intelligent systems and computational tools to accelerate scientific discovery and developer workflows.

Software EngineerMicrobiology Researcher