AI Infrastructure / Open Source SDK
Dart OpenAI
A type-safe OpenAI API client that gives Dart and Flutter teams a production-ready integration surface.
- Ownership
- Created and maintain the library
- Proof
- 666+ GitHub stars
Mohamed Anas Fikhi / GWHYYY
AI Product & Automation Engineer
I turn manual workflows into reliable software. From AI integrations to APIs, SDKs, mobile products, and infrastructure.
666+GitHub stars on Dart OpenAI
5K+Reqistry Play Store downloads
02Verified production engagements
RemoteCasablanca, Morocco
Selected work / 01–06
Systems shipped across AI, products, SDKs, protocols, and mobile.
Range without dilution. Each project shows a different part of the same job: understand complexity, build a usable interface around it, and own delivery.
AI Automation / Event-Driven System
NostrMind
A configurable Nostr monitoring worker that classifies live events with an LLM and can send targeted DM alerts.
- Ownership
- Designed the monitoring and classification flow
- Proof
- Public source code
Production Product Engineering
Gathr
A shipped social product connecting polished cross-platform UX with decentralized Nostr events and backend services.
- Ownership
- Built and shipped the Flutter application
- Proof
- Shipped production application
Fintech / SDK Engineering
Rollee Connect SDK
A Flutter SDK that turns employment-data connectivity into a clean integration surface for third-party applications.
- Ownership
- Sole engineer for the Flutter SDK
- Proof
- Sole SDK ownership
Protocol Engineering / Open Source
Dart Nostr
A Dart library for building typed Nostr clients with relay connections, subscriptions, signing, and broad NIP support.
- Ownership
- Created and maintain the package
- Proof
- 50+ GitHub stars
Mobile Product / Developer Tools
Reqistry
A published mobile product for finding, filtering, and saving public APIs from a developer-focused catalog.
- Ownership
- Designed and built the application independently
- Proof
- 5K+ Play Store downloads
What I build
Production software first. AI, mobile, infrastructure, and automation applied where each creates leverage.
CAP / 01+
AI products & agents
Add model intelligence to real products with explicit tool boundaries and software that handles uncertainty.
- /Tool-using agents
- /Structured extraction
- /Intelligent search
- /Human approval flows
CAP / 02+
Automation & integrations
Connect fragmented tools and replace repetitive handoffs with observable, event-driven workflows.
- /API integrations
- /Webhooks and jobs
- /Internal tools
- /Data processing
CAP / 03+
Product & backend engineering
Take software from first architecture through APIs, application logic, data flow, and production operation.
- /Backend services
- /Production applications
- /Authentication
- /Realtime systems
CAP / 04+
SDKs & developer tools
Turn complex services and protocols into typed, documented interfaces other engineers can trust.
- /SDK architecture
- /API design
- /Packages and libraries
- /Protocol implementations
CAP / 05+
Infrastructure & delivery
Package, deploy, and operate software with practical infrastructure suited to the product and team.
- /Docker
- /Linux and VPS
- /CI/CD
- /Monitoring and deployment
CAP / 06+
Mobile product engineering
Build cross-platform products where mobile delivery is the right interface, backed by strong application architecture.
- /Flutter and Dart
- /iOS and Android
- /Performance
- /Event-driven state
Production AI
Nondeterministic models. Deterministic guardrails.
Useful AI systems need more than prompts. Reliability comes from boundaries, validation, failure handling, permissions, and observable behavior.
R1
Structured outputs and validation
R2
Bounded tools and permissions
R3
Retries and failure handling
R4
Human approval where risk demands it
R5
Observable latency and cost
R6
Deterministic software around model behavior
Tools grouped by what they enable.
AI systems
LLM APIs · Tool calling · Structured outputs · Agents · RAG · Embeddings
Application
TypeScript · JavaScript · Node.js · Dart · Flutter
Data & services
Firebase · REST APIs · Webhooks · Background jobs
Infrastructure
Docker · Linux · VPS deployment · CI/CD
Protocols
WebSockets · Nostr · SDK design · Event-driven systems
Operating model
From unclear problem to operated system.
Architecture earns its complexity. Build the smallest useful slice, verify failure modes, then grow from evidence.
- 01
Understand
Map the real process, constraints, risks, and useful success criteria.
- 02
Architect
Choose the smallest reliable system and define integration boundaries.
- 03
Build
Ship the smallest production slice that proves value.
- 04
Verify
Test failure cases, security, observability, and AI behavior.
- 05
Deploy
Release with monitoring, safeguards, and a clear operating path.
- 06
Iterate
Use real behavior to improve what creates measurable value.
Professional experience
Built and shipped the production mobile application, designed event-driven state, and managed backend services and cloud infrastructure.
- Flutter
- Nostr
- Node.js
- Firebase
- Linux
- Git
Owned delivery of the Rollee Connect Flutter SDK, including its public integration API and developer-facing implementation.
- Flutter
- Dart
- REST APIs
- SDK Design
Open source / Developer ecosystem
Infrastructure other engineers choose to build on.
AI Infrastructure / Open Source SDK
Dart OpenAI
A type-safe OpenAI API client that gives Dart and Flutter teams a production-ready integration surface.
666+ GitHub stars↗
Protocol Engineering / Open SourceDart Nostr
A Dart library for building typed Nostr clients with relay connections, subscriptions, signing, and broad NIP support.
50+ GitHub stars↗
Before we start
Direct answers.
What kind of work is a good fit?
AI product features, workflow automation, API integrations, backend and full-stack product work, SDKs, developer tools, infrastructure-adjacent delivery, and selected Flutter engagements.
Do you work with existing products and codebases?
Yes. I can audit an existing system, map constraints, improve architecture, add integrations, or own a focused product slice without forcing a rewrite.
How do engagements start?
Usually with a short brief and call. From there I define scope, risks, ownership, and the smallest production milestone worth shipping.
Do you only build AI products?
No. AI is one layer in a broader software engineering toolkit. Many useful systems need solid APIs, data flow, product engineering, or automation without model involvement.
Where do you work from?
Casablanca, Morocco. I work remotely with teams and clients worldwide.
Project intake
Have a workflow or product worth improving?
Send context, not a polished specification. I will identify the shortest useful next step.