Work that's live right now

AI Sales Call Analysis & Training Platform

Analyzes sales calls using AI to help agents identify what's working and what's not. Built for sales teams who want data-driven coaching, not gut-feel feedback.

AI Video Generation Platform

Generates 2-3 minute consistent AI videos — something no other tool in the market does currently. Built for content creators and media companies who need production-quality video without production-quality budgets.

AI-Powered Digital Marketing Campaign Tool

Automates digital marketing campaigns using AI-generated video and images. Marketers create full campaign assets in minutes instead of weeks.

Multilingual AI Product Video Generator

Upload a product image, get marketing videos generated in multiple languages automatically. Built for e-commerce brands selling across markets.

A US demi-fine jewellery brand running entirely on Shopify, built and maintained by a partner team since 2020. Six years on, the same team still handles the store, including the work that goes into every sale and festive season.

Ella Stein sells real diamond jewellery starting at around $89, which sounds impossible until you understand where the stones come from. When an engagement ring is cut, a large share of the rough diamond is normally discarded. The brand sorts and uses those offcuts. Everything is made with certified recycled sterling silver and conflict-free stones, and the whole operation is vertically integrated, from design and raw materials through metalwork and handcrafting at their own facility in Mumbai. Since 2017, they've put over 500 women artisans into paid work there. The founder, Adi Choksi, grew up around his family's diamond manufacturing business. The brand is named after his daughter, with "Stein" meaning stone. They sell direct to consumers online and stock through independent jewellers across the US. The Shopify store is the centre of that, and a partner team has owned it since 2020. The full build, the ongoing development, and everything the store has needed since. What makes this one worth reading is the length of it. Six years is a long time to keep a retail client. Sale periods and festive seasons are where jewellery brands make their year, and each one means changes going onto a live store that's taking orders while the work happens. Getting that wrong on the wrong weekend is expensive. The relationship has held through six years of those weekends, and the same team is still handling new work on the account today. Closing line for the card: This is what I mean when I say I know which teams stick around. Six years with one client isn't a case study; it's a track record.

A US direct-to-consumer jewellery brand selling handmade, water-resistant pieces built around colour and a positive-living brand voice. The store runs on Shopify, with a full catalogue, an affiliate programme, and a wholesale channel for retail partners.

Hello Ella sells handmade sustainable jewellery pitched squarely at everyday wear rather than occasion buying. Necklaces, bracelets, studs, hoops, and rings, with pieces starting around $45. The brand voice leans into colour and good-vibe living, which for a jewellery store means the site has to carry personality rather than just display a grid of product photos. The Shopify store handles more than a catalogue. There's a slide-out cart with a free-shipping progress meter that tracks how much more a shopper needs to add, a spend threshold most D2C stores use to lift average order value. Payment coverage runs across cards, Apple Pay, Google Pay, Shop Pay, and Diners Club. Customer accounts are region-aware, resolving to a locale and country at login. Two channels sit alongside the main storefront. An affiliate programme lets creators and partners earn on referral, and a retail partnerships page opens a wholesale route for stockists. A climate commitment page carries the sustainability positioning that the brand is built on. A second D2C jewellery brand on the same delivery team. Two stores in the same niche isn't a coincidence; it's a team that knows the category.

Order Virtually Anything from Anywhere

One of the largest delivery platforms in the region. Its unique on-demand experience has earned the highest user ratings among all major delivery platforms on both the Apple App Store and Google Play Store. With a large network of registered on-demand couriers, it makes this platform possible to order virtually anything from anywhere.

A US hotel furnishings supplier that outfits full-service hotels for owners operating under Best Western, IHG, Choice, Radisson and Hilton flags. The build was awarded end to end to the team I was with, who brought in additional engineers from a partner company to staff it. I coordinated delivery.

The client supplies custom furniture, fixtures, and equipment for hotel renovations, brand conversions, and new construction. Their work covers full room suites, bathrooms and public areas, with custom fabrication alongside major brand prototype furniture. They also handle the paperwork side that hotel owners need for brand approval: technical specifications, drawings, datasheets and mock-ups. The business rebranded from Hospitality Essentials to Modern Hotel Decor during the period the work was running. On delivery: the project came to the team as a full end-to-end engagement rather than a staffing top-up. They pulled in engineers from a partner company to build out capacity, which meant two organisations working on one codebase against one client's timeline. I sat in the coordinator seat across both, running the weekly reporting line to the client and management and keeping the build team aligned on what was moving and what was stuck. This is the kind of engagement that tends to fall apart in the seams, where a client is talking to one company, engineers are sitting in another, and nobody owns the space in between. That's the seat I was in!

A city-wide delivery app that lets users order food, groceries, pharmacy items, and pretty much anything else from local stores, with real-time courier tracking and delivery in minutes. The team I worked with provided dedicated developers who supported the platform's engineering operations on a contract basis.

The platform handles millions of orders across 25+ countries and 1,300+ cities. Users pick from restaurants, supermarkets, pharmacies, and retail shops through a single app. There's also a "send anything" option where someone can request a courier to pick up and deliver items on demand. Three separate panels keep everything running: customers browse and order, courier partners accept and deliver, and merchants manage their storefronts, menus, and inventory. Behind the scenes, the system matches each order to the nearest available courier using real-time location data and estimated delivery windows. The team I was part of provided dedicated in-house developers who worked directly within the platform's engineering workflow on a contract engagement model. This was staff augmentation at scale, with our developers plugging into the client's existing processes, tools, and sprints as if they were internal hires.

A partner team placed senior engineers directly inside the product organisation of a global technology manufacturer. The work covered checkout on their worldwide e-commerce platform, an internal support automation build, and consolidation of their microservices estate.

Enterprise staff augmentation is a different model from project outsourcing, and the difference is where the engineers sit. These weren't handed a specification and left to deliver against it from a distance. They joined the client's daily standups, worked in the client's Jira, pushed to the client's repositories, and were reachable in the client's Slack. No account manager in the middle relaying requirements, no weekly status call standing in for actual collaboration. Three workstreams ran under the engagement. Cart and checkout on a global e-commerce platform. For a manufacturer selling configurable hardware into dozens of markets, this is harder than it sounds. Currency, regional tax rules, market-specific pricing, and product configuration logic all vary by geography, and each one touches the same checkout path. An internal support automation platform, built to cut manual handling out of support workflows that had grown by accretion over years. Microservices consolidation. Large organisations tend to defer this work until the estate becomes difficult to reason about, and then need people who have unpicked one before. Closing line for the card: I'm not claiming this as my deal. It's here because it shows what the teams I work with can achieve when the client operates at that scale.

A partner team was brought in to modernise the platform behind a large international education provider. The work covered upgrading a set of .NET microservices, moving the data layer to GraphQL, widening single sign-on, and laying the groundwork for AI features the platform didn't yet support.

Education platforms at this scale tend to accumulate. Services get written, requirements shift, and a few years later the estate works but nobody wants to touch it. This engagement was about making it workable again rather than replacing it. Ten .NET microservices were upgraded. The kind of groundwork that unblocks everything downstream and shows up in no release note. The data layer moved onto GraphQL. For a platform serving many kinds of client applications, that means the front end asks for what it needs instead of pulling whole payloads and discarding most of them. Single sign-on was widened to cover Google and Microsoft accounts. This sounds routine until you consider that education platforms sit across institutions, corporate learners, and individuals, all with different identity providers behind them. Infrastructure preparation for AI features. Not the features themselves, but the plumbing that has to exist first. This is the work that decides whether an AI roadmap takes months or years, and it's almost always invisible from the outside. Engineers worked inside the client's own environment and process rather than delivering to a specification from a distance. Closing line for the card: These are engagements my delivery partners have run. I'm not claiming them as my deals. They're here so a founder can see the level my partners operate at when the client is a large organisation with a decade of legacy behind them.

Enterprise Client: A partner team ran two workstreams for a global design software company. The first automated the connection between their cloud platform and their desktop applications used in architecture and civil engineering. The second built out sustainability tooling for carbon analysis.

Design software for buildings and infrastructure sits in an awkward place. The heavy work happens in desktop applications that engineers have used for decades, while everything around it has moved to the cloud. Getting those two worlds to talk to each other properly is an engineering problem in its own right, and it's the first thing this engagement dealt with. Cloud automation between platform and desktop. The applications involved handle building information modelling and civil infrastructure design, which means large, complex files with dependencies that break easily. Automating the flow between desktop and cloud without corrupting model integrity is the hard part, and it's why this work tends to need people who understand the file formats rather than generalists. Sustainability and carbon analysis tooling. Construction and the built environment account for a substantial share of global emissions, and regulation increasingly requires that impact be calculated at design stage rather than estimated afterwards. This workstream extended the platform's ability to run that analysis, putting carbon numbers in front of engineers while decisions are still cheap to change. Engineers embedded within the client's development workflow rather than delivering to a specification from outside. Closing line for the card: I'm not claiming this as my deal. It's here to show the range my delivery partners work across, from commerce to legacy platforms to specialist engineering software.

A fitness app had thousands of people logging workouts and nothing useful to tell them about it. A partner team built an AI layer on top of the existing product that reads workout history and biometric trends, then returns personalised insights and adaptive next steps.

The app worked. People used it, logged their sessions, and came back. What it couldn't do was say anything meaningful about what it was collecting. Users saw their numbers played back at them, and the coaching was the same for everyone regardless of what the data showed. Retention had flattened out, which is what usually happens once the novelty of tracking wears off and nothing replaces it. The fix wasn't a rebuild. A partner team layered AI analytics onto the product as it stood, reading workout history, biometric trends, and session patterns to work out what was actually happening with each person. Out of that came individual performance insights and next-step recommendations that adapt as someone progresses, rather than a fixed programme that ignores whether they're improving or struggling. All of it surfaced inside the existing app. No migration, no parallel product, nothing users had to relearn. The team reported a 38% rise in session engagement, with early-stage churn dropping within two months of launch. Closing line for the card: This is the kind of work founders come to me for. The product already exists and already has users. What's missing is the intelligence layer that makes the data worth something.

A WordPress site built for a London-based psychoanalyst and organisational consultant who wanted a permanent home for his professional writing, talks, and papers. I sourced the client, brought in a technical lead to scope the requirements, and passed delivery to GrowIdeas Innovations LLP, who designed and built the site.

The client had decades of published papers, radio talks, and lecture material scattered across old files and had decided to bring it all into one place while he wound down his consulting practice. He needed something he could update himself without calling a developer every time. The build covers four content types on one WordPress install: short-form blogs sorted into social, political, professional, and philosophical categories, downloadable PDF papers, embedded video talks with slide decks, and a contact form for reader responses. Category filtering runs across the blog and video sections so visitors can navigate by theme instead of scrolling chronologically. Built on Elementor so the client edits and publishes on his own. I found this lead through LinkedIn outreach in February 2025, set up the initial call, and brought in a technical lead to walk through requirements with the client directly. The team at GrowIdeas Innovations handled design and development. I stayed involved through delivery.

A UK corporate wellbeing app covering mental and physical fitness, sold to businesses rather than individuals. A partner team built the app and has been on it since 2021, with two developers still assigned full time today.

Revoola is a London FitTech company founded in 2018, with the app launching in January 2020. The pitch to a business isn't fitness, it's absenteeism. Founders Mark Williams and Paul Goossens built the product around the argument that employee mental and physical health shows up in productivity and sick days, which turns wellbeing into something a finance director can sign off on. The app carries both halves of that. Instructor-led mindfulness sessions and body fitness classes stream live or on demand, covering running, indoor cycling, HIIT, yoga, pilates, and flexibility work. Open workouts let people log their own running, walking, or riding with distance, time, heart rate, calories, and elevation attached, so activity outside the app still counts toward what an employer sees. It syncs with Wear OS in real time. The social layer is what separates it from a video library. Interactive leaderboards compare performance live regardless of where people are, and any class can be scheduled as a group session or a one-to-one challenge. For a workforce spread across home and office, that turns a solitary activity into something a team does together. A partner team built the application and has stayed on it since 2021. It runs cross-platform on Flutter, with a dedicated iOS developer working alongside the Flutter developer. Five years later, both are still assigned full-time, and the work is ongoing rather than in maintenance. Google Play still shows active releases. That last part is the point worth reading. A wellbeing app sold on a B2B contract has to keep working every day for an employer who is measuring it. Five years of continuous development with a permanently assigned team is what that actually requires, and it's a different commitment from shipping a build and moving on. Delivered by a partner team. Still delivered by them, five years on and still running.

A web-based operations system for a road freight company, covering consignment paperwork, fleet assignment, freight billing, and party accounts in one place. Built by the team at NovusCode. I handled client coordination, calls, and invoicing on the engagement.

Indian road freight runs on the builty, the consignment note that travels with every load and records what was picked up, who sent it, where it's headed, what's owed, and on what terms. Most operators handle these on carbon-copy books with a separate ledger for payments. Nothing is searchable, nothing reconciles on its own, and one misfiled document can stall a payment for weeks. The system replaced that with a single web application covering the full operational cycle. Consignments are generated and printed in the format drivers and consignees expect. Each one gets a vehicle and driver assigned against it, with trip records held alongside. Freight charges calculate against the operator's own rate structure rather than a fixed table, which matters because transport pricing shifts by route, load type, and party. Billing runs GST-compliant invoices off the same data, so the paperwork and the accounting draw from one source instead of two. The receivables side is where the real value sits. Party ledgers track what each client owes and what's outstanding, which for a transporter is usually the difference between a good month and a cash crunch. Delivery status closes the loop, so the office can answer where a load is without ringing a driver. I worked on this during my time at NovusCode. My side was the client-facing layer: running calls, keeping requirements moving between the client and the build team, and handling invoicing on the engagement.