Open to AI Engineer & ML Engineer roles

Field manual / 2026

Models answer.I engineer what happens next.

01 Context: Screen · voice · text
02 Retrieval: Rank · bound · trace
03 Routing: Language · vision · speech
04 Evaluation: Quality · latency · failure

Flagship evidence Volyx Lens Inspect the system

I'm Joshua, a Python-first AI Engineer building the retrieval, multimodal input, model routing, backend reliability, and evaluation layers that turn model capability into useful systems.

Inspect the evidence View CV

Interactive system field The model is one node. The engineering is the field.

context plane control plane

01 / context

Select relevant text, screen, voice, and meeting context before a model sees it.

Python backends

RAG + context engineering

Multimodal + voice AI

Evaluation before claims

Selected evidence / 05

Not a wall of logos.
Work you can inspect.

Five builds showing context retrieval, model integration, async APIs, realtime state, and honest evaluation. Every claim has somewhere to go.

Direct AI evidence

Multimodal AIContext engineering

Volyx Lens

A privacy-oriented macOS assistant combining selected screen context, microphone input, meeting audio, local OCR, relevance-ranked retrieval, transcription, and multiple AI providers. I hardened its context boundaries, provider routing, consent controls, tests, and release checks.

Screen, speech, and meeting context
Ranked retrieval with explicit boundaries
Provider routing and consent controls

Supporting engineering

Real-time systemsBackend engineering

Noughtline

A React and Express multiplayer game with signed guest sessions, server-validated moves, reconnect handling, an append-only transaction ledger, and one-time crediting for verified Paystack references.

Python integration

Async API integrationPython

VirusTotal Telegram Bot

An async Telegram bot for scanning files, URLs, and hashes through VirusTotal v3. It hashes locally before upload, follows the large-file upload flow, backs off on rate limits, and caches reports for interactive details.

Flow Hash locally → query → cache report

Python integration

API aggregation prototypePython

Solana & Ethereum Wallet Analyzer

An async Telegram bot that validates Solana and Ethereum addresses, combines Solana and Ethereum RPC, CoinGecko, and DexScreener data, caches requests for five minutes, and paginates Solana token holdings in chat.

Flow Validate → gather → merge → cache

Archived experiment

ML evaluationPython

Football Predictor

An archived forecasting experiment combining calibrated XGBoost and Poisson score probabilities. Its real-data evaluation demonstrated signal over naive baselines but not practical value over bookmaker probabilities.

Rolling-origin evaluation across 1,140 Premier League test matches
53.77% accuracy versus a 56.70% bookmaker benchmark; not a betting product

Conclusion Evidence changed the product decision

Operating method / building VolyxAI

Capability is cheap.
Control is engineered.

Building VolyxAI is one place I apply this method: language, vision, and speech models connected to relevant context, structured outputs, reliable services, and human review where consequences matter.

01
Integrate
Connect language, vision, and speech providers behind explicit interfaces.
provider layer
02
Retrieve
Assemble ranked, bounded, source-aware context for the task.
context layer
03
Validate
Use schemas, deterministic checks, retries, and human approval.
control layer
04
Evaluate
Measure quality, latency, cost, and failure behavior before claims.
evidence layer

See what I'm building at VolyxAI

Engineer profile

Past the prompt.
Into the system.

I build Python-based systems around language, vision, and speech models: provider integrations, RAG and context assembly, FastAPI services, asynchronous workflows, structured outputs, and evaluation.

I follow failures into APIs, webhooks, databases, queues, authentication, retries, and the edge cases between them. The work should be inspectable, operable, and honest about what the evidence supports.

AI engineering: RAG · context engineering · multimodal · voice · structured outputs
Models & providers: Azure AI Foundry · OpenAI · Anthropic · Gemini · DeepSeek · Deepgram
Languages: Python · SQL
Backend & evaluation: FastAPI · asyncio · REST · n8n · scikit-learn · XGBoost

Open channel

Need an engineer for the part after “call the model”?

Open to AI Engineer and ML Engineer opportunities.

Discuss a role CV LinkedIn GitHub