Singularity Shift Ltd
Start a project
AI-assisted portrait of James Walford in his software studio

Founder-led software engineering · Applied AI · Cheltenham · UK-wide

AI is changing what software can do. I make it dependable.

I’m James Walford, founder of Singularity Shift Ltd. I help individuals, founders and teams decide where AI belongs, then shape web, mobile, desktop and proof-oriented software people can use, understand and trust.

Describe the problem See selected work

No project too small · Individuals to larger organisations · Work directly with James

01Bring a product to life 02Replace a fragile process 03Rescue an existing system 04Build for real devices 05Make trust verifiable

Selected work

Products and systems where trust matters.

I build across the product people touch and the difficult systems underneath it—carrying work through AI-assisted workflows, architecture, implementation, testing and real-world delivery.

Product platform

Plyapp.io

Built and improved a production marketplace platform spanning a React frontend, Rust API, PostgreSQL data and customer-facing email workflows.

  • React
  • Rust
  • PostgreSQL

Mobile product

Nova Wallet

Designed and built a non-custodial Android wallet for Move networks, taking the product through real-device delivery and Google Play Early Access.

  • Android
  • Move
  • TypeScript

Desktop software

Nova Desk

A cross-platform desktop wallet for the Cedra network, with encrypted storage, DApp integration and an update path designed for real-world releases.

  • Rust
  • Desktop
  • Cross-platform

Cryptographic systems

zk-proofs

Built proof-oriented infrastructure for signed, portable records that can be verified offline and selectively disclosed when the underlying information should stay private.

  • Proof systems
  • Cryptography
  • Selective disclosure

Applied AI

The age of AI needs more than a chatbot.

AI is changing what software can do. The useful question is where it creates leverage without creating a new source of confusion, risk or unreviewable decisions.

AI strategy

Find the leverage

Map the actual work, data and decisions first. Identify where AI can remove friction, improve access to knowledge or make a product more capable.

  • Opportunity mapping
  • AI strategy

AI engineering

Build the surrounding system

Connect models and tools to the interfaces, permissions, data, fallbacks and workflows that make them useful in practice.

  • AI workflows
  • Knowledge tools

Trust & evaluation

Make it accountable

Use evaluation, evidence, human review and clear failure paths so people know what the system did and when it is safe to rely on it.

  • Evaluation
  • Human oversight

Ways I can help

Build what AI makes possible. Keep it dependable.

AI creates possibility quickly. The work is deciding where it genuinely helps, what to augment, what to automate and what must stay human. Then we build the smallest dependable system around that decision.

01

Products people can use

Shape and build web, mobile or desktop products from the first useful decision through testing, release and the next sensible phase.

02

AI systems, workflows & automation

Connect products, data and AI-assisted operational workflows so teams spend less time fighting disconnected tools.

03

Mobile & desktop software

Build for real devices and workflows, with careful attention to performance, failure handling, release delivery and the person using it.

04

Proof systems & decentralised engineering

Build Move modules, wallet integrations and proof-oriented infrastructure where correctness, verifiability and clear user experience matter.

05

Rescue, modernise & integrate

Untangle a stalled build, stabilise software that has become difficult to change, or connect systems that were never designed to work together.

A practical toolkit, chosen around the work

Rust TypeScript React Node.js Python PostgreSQL Docker Move AI systems

How the work moves

Clarity before complexity. Evidence before assumptions.

Good software begins with a clearer decision, not a longer feature list.

I surface the goals, constraints and risks early, then turn them into an achievable piece of work whose progress and trade-offs you can understand.

  1. 01Find the real problem

    Clarify the outcome, the users, the existing system and what is getting in the way.

  2. 02Choose the first move

    Reduce uncertainty and make the valuable first phase, risks and trade-offs explicit.

  3. 03Build visibly

    Deliver working progress, test assumptions and explain important decisions as we go.

  4. 04Ship with clarity

    Launch with the testing, documentation and understanding needed to extend the work.

How projects start

A sensible first move, shaped around the work.

Whether you are an individual with a useful idea or a larger organisation with a difficult system, the starting point depends on the work—not the size of your organisation.

For a focused need

One useful change

Start with one awkward workflow, useful integration, prototype or contained improvement—scoped to the result you need now.

For a defined workstream

A product or system to own

Bring in specialist ownership for a product build, difficult integration, desktop/mobile software or technical rescue alongside your team.

Bespoke does not mean vague.

Every engagement is priced individually. After an initial conversation, I will recommend the most sensible next step and set out the proposed scope, assumptions, delivery shape and cost before the build begins. If the unknowns are still too large, we can start with a clearly defined discovery phase.

You pay for the problem we are solving—not a package filled with things you do not need.

Tell me what you need A rough outline is enough. No finished brief or chosen technology required.

James Walford · Founder, full-stack developer & AI practitioner

I understand the AI stack—and still care about the person pressing the button.

I work in the difficult middle between an AI possibility and a dependable product.

My route into software has been hands-on: building, breaking, investigating and rebuilding complete products. It taught me to move from an interface problem to the APIs, data, infrastructure and product decisions behind it—and to recognise when the surrounding design, rather than the immediate symptom, is the real problem.

My AI work is practical and whole-system: models are only one part. I think about the data they can use, the actions they can take, how outputs are evaluated and where human judgement must remain.

With Singularity Shift, you work directly with the person scoping the system, making the technical decisions, writing the code, testing the behaviour and explaining the work.

  1. 01Whole-system thinking

    Interface, services, data, infrastructure and delivery treated as one connected product.

  2. 02Practical judgement

    Technology, model or workflow chosen for the problem, its constraints and the burden it creates later.

  3. 03Ownership through launch

    A feature must behave correctly, survive failure and make sense to the people using it.

  4. 04Built to be understood

    Clear code, evaluation and explanation so the work can keep earning its place.

Inferenco blue flame logo

Product partnership

Broader collaboration when the work calls for it.

When a project needs broader collaboration, I can work through the Inferenco partnership across applied AI, mobile and decentralised technology.

Visit inferenco.com

A rough outline is enough

What needs to work better?

Tell me what you are trying to achieve, what is getting in the way and where things stand today. It can be a new idea, an existing system or a problem you cannot yet define technically—including a new AI capability or intelligent automation.

Email James about your project

No project too small · Individuals to larger organisations · No finished brief required