AI MUSIC / AUDIO
MuseMify
An AI music and audio creator for song generation, covers, monologues, audio extension, lyrics, and social discovery.
View on Google Play ↗Available for AI Engineering & Flutter roles
Products that make
technology feel human.
I'm Marcus Jacob, a Flutter developer and AI application engineer building polished mobile products, intelligent experiences, and the systems behind them.
Currently building AI-powered applications at Cyrex Ltd.
Port Harcourt, Nigeria · Open to remote
01 / 04Scroll to explore ↓
20K+combined app downloads
4.5+average product ratings
FlutterAndroid · iOS · Web
AI-nativeLLMs · RAG · agents
02 / Selected work
Built for real people.
Shipped into the real world.
MULTI-MODAL AI
ImageFlux
A mobile creative AI studio for generating images, videos, music, voiceovers, and lip-synced content across multiple models.
View on App Store ↗RESEARCH / RAG
◌◌◌
MediCortex
An AI research assistant over 50,000+ PubMed papers with hybrid retrieval and citation-enforced answers.
Explore the repository ↗CONVERSATIONAL AI
✦
Ask better.
Build faster.
Build faster.
ChatO1 + MJ Chatbot
Consumer AI experiences with multi-provider model access, web search, image generation, voice mode, and specialist personas.
See more on GitHub ↗Professional delivery / Cyrex Ltd
More products,
one reliable standard.
Alongside my independent AI products, I’ve contributed to client-facing work at Cyrex across hospitality, payments, logistics, and consumer platforms.
See the Cyrex project archive ↗OPEN SOURCE / FLUTTER
⌘
Store<State> → scoped rebuilds → testable effects
03 / Developer tool
StateForge
Flutter state, forged for features.
An open-source, feature-scoped state-management toolkit for Flutter with typed stores, selective rebuilds, one-time effects, persistence, undo/redo, and pure Dart testing utilities.
04 / Engineering evidence
The systems behind
the product experience.
Architecture snapshots from the codebases behind the work: the boundaries, reliability decisions, and developer tooling that make the product layer possible.
01 / MEDICORTEXRAG SYSTEM
Grounded medical research answers
Flutter workspace↓
source-bound citations
Vertex AI
768D query embeddingQdrant
vector + text retrieval
↓Gemini answer768D query embeddingQdrant
vector + text retrieval
source-bound citations
50K+ target papersSearch-time telemetry
Separates search, AI, MCP, conversation, analytics, and settings domains. The RAG service builds a source-bound context, enforces citations, returns confidence, and records search time.
Role — Sole code contributor; architecture & implementation.
Explore architecture ↗02 / CHATO1MODEL ORCHESTRATION
AI providers as a runtime capability layer
Flutter client↓Backend model catalog
runtime discovery↓
runtime discovery↓
Provider router
streaming + attachmentsTool manager
web · news · image · code · files
streaming + attachmentsTool manager
web · news · image · code · files
5 built-in toolsDynamic provider catalog
Uses a provider contract for model capabilities, connection testing, streaming, and disposal. The client discovers backend model metadata at runtime, then routes requests and tool calls through a shared service boundary.
Role — Sole code contributor; architecture & implementation.
View product ↗03 / MUSEMIFYMEDIA WORKFLOW
Generation tasks that survive the session
Flutter + Riverpod↓Task repository
20-item paged history↓
20-item paged history↓
Appwrite data + storage
task, audio, artwork lifecycleAudio service + cache
sessions, downloads, offline
task, audio, artwork lifecycleAudio service + cache
sessions, downloads, offline
Paged task workflowOffline playback
Feature, data, domain, and presentation layers keep generation, playback, library, social, subscription, and offline work independently evolvable. Media playback handles sessions, notifications, buffering, playlists, and local availability.
Role — Sole code contributor; architecture & implementation.
View product ↗04 / STATEFORGEDEVELOPER TOOLING
State infrastructure with observable behaviour
Store<State>↓
contracts + stress benchmarks
Scoped rebuilds
selectors + microtasksEffect stream
one-time events
↓DevTools + testsselectors + microtasksEffect stream
one-time events
contracts + stress benchmarks
Pure Dart coreBenchmark suite
The package is split into a pure Dart core, Flutter bindings, a shared-preferences adapter, DevTools extension, example app, and benchmark suite—so runtime behaviour can be tested, profiled, and inspected.
Role — Sole code contributor; architecture & implementation.
View on pub.dev ↗Architecture details shown here are intentionally high-level and drawn from my own project implementations; client-specific internals remain confidential.
05 / Engineering approach
Strong fundamentals.
Curious at the edges.
01
Build for the whole product
I care about the experience from first tap to shipped release: thoughtful UI, resilient state, useful APIs, and the product decisions that connect them.
02
Use AI with intent
I work across OpenAI, Claude, Gemini, and Grok, designing multi-provider systems, RAG pipelines, agent handoffs, and generative workflows that solve a real user problem.
03
Make complexity legible
Clean architecture, typed boundaries, predictable state, and automation make fast-moving teams calmer—and make products easier to evolve.
Core stack
Languages
Dart · Python · TypeScript
Mobile
Flutter · Riverpod · Android · iOS
AI systems
RAG · MCP · LLM orchestration · Prompt engineering
Infrastructure
Firebase · Appwrite · PostgreSQL · AWS / GCP · CI/CD
06 / A little about me
Engineer by craft.
Builder by instinct.
I enjoy the intersection of mobile engineering and applied AI: taking a rough idea, giving it a reliable shape, and shipping something people choose to keep using.
At Cyrex Ltd, I work on AI-powered consumer and enterprise products. Outside the editor, you’ll usually find me making music, thinking about interfaces, or exploring a new tool just to see what it can do.
Let's build something useful