Software Engineering
Web platforms built end to end: React and Node in front, Django, Laravel, and Express behind, wired to the databases and services underneath.
Emmanuel DavidSoftware Engineer Machine Learning AI Creative
I build software that holds up: machine learning systems, applied AI, and the visual and interaction design that makes them feel considered rather than merely functional.
Most of my work sits where engineering and craft overlap: shipping an interface, then the models and infrastructure behind it.
I'm a Software Engineer working across machine learning, applied AI, and creative practice. Most of what I build sits at the seam between those: systems that are technically sound and still feel like someone bothered.
In practice that means training and shipping models, wiring them into real applications, and designing the interfaces and identities that carry them. I care about the unglamorous parts just as much: latency, authentication, deployment, and the code somebody else has to read next year.
I like problems that refuse to respect a single discipline. A feature might need a model, a design system, and an API contract inside the same week, and I like being the person who doesn't have to hand that off.
Outside client work I photograph, shoot film, and keep a visual practice running. It's the same instinct as the engineering: pay attention, then strip out everything that doesn't earn its place.

One person carrying the whole thing: engineering, modelling, and visual craft, without the hand-off in the middle.
Web platforms built end to end: React and Node in front, Django, Laravel, and Express behind, wired to the databases and services underneath.
Supervised and unsupervised models, deep learning, and computer vision with TensorFlow and scikit-learn, containerised so inference is reproducible.
Retrieval, agents, and LLM-backed features moved out of the notebook and into production, with evaluation, guardrails, and cost that stays predictable.
Photography, film, and visual identity: art direction, type, and design systems made to be used rather than admired and left behind.
A selection across applied AI, software infrastructure, fintech, Web3, and visual work: each one something I took from idea to shipped.
Retrieval-augmented generation platform for working through large document sets: grounded answers, citations, and evaluation on every query.
Autonomous agents combining computer vision and vector search to analyse market data and surface the insight worth acting on.
Web3 platform with wallet authentication, blockchain infrastructure, referral systems, and an interface that stays fast as it grows.
A financial assistant that tracks activity and explains it in plain language, so on-chain behaviour is legible instead of cryptic.
Machine-learning detection for live network traffic: anomaly scoring and threat analysis at the pace real systems produce events.
Ongoing photography, film, and brand work: portraiture, product imagery, and identity systems built to be used rather than just admired.