Software Engineer

I build full-stack, backend, and production AI systems people can actually use.

I'm Mayowa, a Software Engineer building autonomous agents, forecasting pipelines, backend services, and production interfaces. I choose between deterministic and agentic approaches based on the problem, then build the validation, recovery, and product layer that make the system dependable.

View selected work

Agentic AIML pipelinesFull-stack systemsRetail intelligenceAWS / GCPReact + Node.js

Based in GMT+1 and hybrid/remote-friendly.

Open to full-time, contract, and high-ownership AI product work.

Open to teams building reliable AI products with real users.

HeySynth operations cockpit showing retail signals, inventory risk, and suggested actions

HeySynth / Product proof

Featured Build / HeySynth

Building forecasting, agent, and operator workflows for retail teams managing inventory decisions.

Role: Full-Stack AI/ML Engineer working across ML services, backend workers, main service APIs, agent systems, and operator-facing product surfaces.

Forecasting engineWorker pipelinesAgent orchestrationBackend APIsReact surfaces

At HeySynth, I work across the parts of AI products that usually get split across teams: forecasting logic, worker pipelines, agent orchestration, service APIs, and the React interfaces operators use to act on AI recommendations.

Productionized forecasting paths for uneven product histories: tiered model selection, enhanced baselines, retail-specific forecast services, and forecast-risk modes.

Built backend and worker flows for spreadsheet ingestion, forecast imports, OTB processing, validation, and realtime completion/error updates.

Hardened agent workflows for demand planning, retail intelligence, and forecasting by improving tool contracts, prompt guardrails, routing, and output safety.

Shipped React surfaces across forecasting, agent, and collaboration workflows so operators could understand and act on AI output.

View full case study See all work

Engineering judgment

I choose the simplest reliable path for the problem.

AI is not automatically the answer. I decide where deterministic software is stronger, where agentic behavior adds value, and what controls make either path dependable in production.

Deterministic systems

Use explicit workflows when the path is knowable.

For stable rules, predictable state transitions, and high-consequence operations, I prefer clear contracts, validation, retries, and recoverable failure paths.

Agentic systems

Use agents when the work needs judgment and adaptation.

For open-ended tasks, tool selection, and natural-language interaction, I add bounded autonomy with state, approvals, observability, and safe terminal outcomes.

The engineering work is in the boundary: validation, state, recovery, observability, and a product interface that makes system behavior understandable to the people using it.

Build Loop

I turn unclear business problems into shipped AI systems with a practical build loop.

STEP 01

Problem Scoping

Identifying where AI can reliably improve a real business KPI, not just where it sounds impressive.

STEP 02

Architecture Design

Defining system boundaries, failure paths, and data contracts before implementation.

STEP 03

Data Pipeline

Building ingestion and validation first, because model quality depends on data quality.

STEP 04

Model Development

Training and evaluating models against business-relevant metrics, not benchmark vanity.

STEP 05

Agent Orchestration

Wiring tools, memory, and routing so agents stay useful, grounded, and controllable.

STEP 06

Full-stack integration

Shipping the intelligence layer through APIs and UI workflows people can actually use.

Experience

Work shaped by shipping real systems

A timeline of roles where I shipped production AI systems across agent orchestration, forecasting, full-stack integration, and MLOps reliability.

Jul 2025 - Present

Full-Stack AI/ML Engineer

HeySynth · Techstars-backed · Remote (Austin, USA)

Architecting multi-agent retail systems, from forecasting logic and agent orchestration to production-ready frontend delivery.

Built LangGraph-based multi-agent systems for retail planning workflows.

Shipped Python forecasting services with tiered model routing and sparse-data fallbacks.

Implemented FastAPI + Pub/Sub async pipelines with human-in-the-loop validation.

LangGraphMCPFastAPIGoogle Pub/SubVertex AIReact/TypeScript

Apr 2025 - Jul 2025

Full-Stack AI/ML Engineer

InspireEdge · Remote (UK)

Delivered production AI modules for e-commerce SMEs, including market intelligence, risk detection, and explainable recommendation workflows.

Built end-to-end LLM and RAG workflows for explainable merchant insights and recommendations.

Integrated AI services with Shopify and WooCommerce dashboards for direct in-product intelligence.

Optimized cart-abandonment and anomaly pipelines with model monitoring in Weights & Biases.

RAGPrompt EngineeringFastAPIShopifyWooCommerceW&B

Jan 2024 - Mar 2025

AI/ML Engineer

Nabafat.AI Technologies · Remote (California, USA)

Built enterprise multi-agent systems, multi-LLM orchestration pipelines, and production AI applications across healthcare, cybersecurity, and developer tooling.

Built enterprise multi-agent workflows for startup advisory and personalized learning use cases.

Engineered multi-LLM routing across GPT-4, Llama 3.1, and DeepSeek for better cost-quality tradeoffs.

Shipped AI products including an MRI tumor classifier and a phishing-detection API.

LangChainOpenAILlama 3.1DeepSeekTensorFlow LiteDocker

Mar 2023 - Aug 2023

AI/ML Engineer

Quantum Leap Limited · Remote

Improved ML training economics and reliability for business forecasting and decision-support systems.

Optimized training and inference paths for forecasting and decision-support systems.

Built sentiment and churn models to support data-driven product decisions.

Implemented MLFlow experiment tracking and data versioning for reproducible training cycles.

TensorFlowJAXNumPyMLFlowForecastingMLOps

Jan 2022 - Feb 2023

Data Scientist, Contract

iBloov

Delivered contract data science and analytics work to support practical business and product decisions.

Data ScienceAnalyticsMachine Learning

Nov 2019 - Dec 2021

Data Scientist, Contract

PKVEST

Delivered contract data science and analytics work to support practical business and product decisions.

Data ScienceAnalyticsMachine Learning

AI Labs

Prototype Builds

Smaller builds where I explore realtime AI, agent orchestration, retrieval, and interaction patterns. These are not flagship production case studies, but they show how I prototype, test ideas, and ship with modern AI tools.

Demo build

Vocalis AI

Real-time voice assistant for continuous conversations, with live waveform visualization, interruptibility, and emotion detection using OpenAI Realtime API and WebSockets.

Next.js

OpenAI Realtime API

WebSockets

React Audio Viz

Edge Functions

Open demo

Prototype

AgentGraph Studio

A fully functional visual orchestrator for multi-agent LLM workflows. Features interactive "Agent Nodes" with customizable system prompts (e.g., "Sarcastic Poet"), real-time model switching (GPT-4 vs GPT-3.5), and live execution visualization with feedback loops.

React Flow

OpenAI API

Zustand

Shadcn UI

Next.js

Open demo

View flagship case study

Peer Signals

What collaborators notice after building with me

Mayowa became both a trusted teammate and a confidant while we built from zero to launch.

We started at stage zero, and he helped us get all the way to distribution on both the Apple App Store and Google Play. He is a gem and a wonderful asset to any team or project.

Michael Morgan Lawrence

He raises product quality without slowing delivery.

Across product and engineering conversations, Mayowa turns rough ideas into clear execution plans and reliable shipping momentum. He is practical, fast, and consistent under pressure.

Brandon Orand

Mayowa is deeply capable, collaborative, and dependable on hard builds.

He communicates clearly, takes ownership of ambiguous problems, and helps the whole team move faster without sacrificing quality. Working with him makes complex delivery feel manageable.

Jerome Tullo

He bridges AI architecture and real operator workflows.

Mayowa can move from system design to implementation cleanly. That reduced handoff friction across product and engineering and helped us ship features that were both technically strong and usable.

Victor Essang

Strong product thinking and strong execution in the same person.

Mayowa balances user impact with technical depth. He asks the right questions early, aligns stakeholders quickly, and follows through with implementation quality that teams can trust.

Rachael Oyelami

Notes

Writing from the build floor

Notes, technical essays, and older field logs from building with code, data, and AI systems.

View Archive

Mon, 24 Jul 2023 21:06:48 GMT

Introduction to ReactJS: A Beginner’s Guide

A beginner-friendly guide to React fundamentals: components, JSX, props, state, and how the framework helps structure interactive user interfaces.

Tech

Article

Fri, 07 Jan 2022 19:14:13 GMT

EULER: SIMPLY THE BEST

A DeFi explainer on Euler Finance, permissionless lending, asset tiers, reactive interest rates, protected collateral, and how blockchain protocols rethink borrowing markets.

Tech

Article

Sat, 18 Sep 2021 20:09:07 GMT

The Real ‘Tech Bro’ Introduction Yaayyy!

An archive note from my earlier writing, kept here as part of the public trail of learning, building, and thinking through ideas over time.

Tech

Article

View Archive

Start a conversation

Have an AI product that needs to work in the real world?

Send me the context: what you're building, what already exists, and where the hard part is. I'm especially useful when the work cuts across model, backend, and product.

Best fit for

Agentic AI products that need to leave demo mode

ML pipelines, forecasting, and data-heavy backend work

Full-stack systems where the model, API, and UI need one owner

hello@mayowaadeoni.comHello talks@mayowaadeoni.comTalks work@mayowaadeoni.comWork linkedin.com/in/mayowa-adeoniLinkedIn @_maycodeX

I build full-stack, backend, and production AI systems people can actually use.

Building forecasting, agent, and operator workflows for retail teams managing inventory decisions.

I choose the simplest reliable path for the problem.

Use explicit workflows when the path is knowable.

Use agents when the work needs judgment and adaptation.

I turn unclear business problems into shipped AI systems with a practical build loop.

Problem Scoping

Architecture Design

Data Pipeline

Model Development

Agent Orchestration

Full-stack integration

Work shaped by shipping real systems

Full-Stack AI/ML Engineer

Full-Stack AI/ML Engineer

AI/ML Engineer

AI/ML Engineer

Data Scientist, Contract

Data Scientist, Contract

Prototype Builds

Vocalis AI

AgentGraph Studio

What collaborators notice after building with me

Mayowa became both a trusted teammate and a confidant while we built from zero to launch.

He raises product quality without slowing delivery.

Mayowa is deeply capable, collaborative, and dependable on hard builds.

He bridges AI architecture and real operator workflows.

Strong product thinking and strong execution in the same person.

Writing from the build floor

Introduction to ReactJS: A Beginner’s Guide

EULER: SIMPLY THE BEST

The Real ‘Tech Bro’ Introduction Yaayyy!

Have an AI product that needs to work in the real world?

Let's build useful AI systems, not impressive demos.