I reduced maintenance-record lookup time by 35% for a facilities team by building a Python retrieval assistant linked to its engineering manuals.
Find a Freelancer
302 freelancers ready to help
I develop AI knowledge assistants with the final users and operating constraints in mind. I focus on a concrete deliverable, the people who will use it, and a handover that makes the next iteration manageable. Teams receive the completed work with the key assumptions and testing steps documented, so they can maintain it after delivery.
I improve visual materials for product teams by removing manual steps and unclear ownership. Rather than broad recommendations, I leave a usable process, working files, and clear next steps for the team. Teams receive the completed work with the key assumptions and testing steps documented, so they can maintain it after delivery.
I build controlled automation for support teams for client success teams. I work directly with the people who approve and use the result, which keeps changes tied to the actual operational need. Teams receive the completed work with the key assumptions and testing steps documented, so they can maintain it after delivery.
I help building maintenance teams replace disconnected spreadsheets with a clear planning process. When managers cannot compare work requests and capacity, I map the handoffs and build a manageable reporting structure. My work includes analysing maintenance backlogs, interviewing supervisors and creating Excel planning models. I provide decision criteria and an implementation plan that fits the team’s existing systems.
I developed a web-based archive assistant for an environmental charity after running a small software studio focused on public information. The application combines a Python retrieval service with a React interface for staff reviewing source documents. I define an evaluation set with the client, connect approved repositories and test whether each answer links back to relevant evidence. Access controls are configured around the organisation's existing user roles. My current practice includes maintaining the service, reviewing updates and preparing deployment notes for the client's technical team.
I create retrieval and evaluation workflows for teams that need to make decisions from reliable information. Rather than broad recommendations, I leave a usable process, working files, and clear next steps for the team. This approach keeps the final result grounded in what the team can verify and continue using.
I built a rolling cash model for a family-owned greenhouse that needed to plan seed purchases ahead of seasonal sales. The model separates growing costs, customer receipts and payroll so the owner can test different order volumes. I prepare cash forecasts, management accounts and practical spreadsheet controls for independent growers and distributors. My work uses Excel and bookkeeping exports, with documented assumptions and checks that the client can update.
I reconcile unreliable cash-flow forecasts for small manufacturers, using Excel to investigate missing entries and protect sensitive financial data.
I develop editorial systems and terminology guides with the final users and operating constraints in mind. I focus on a concrete deliverable, the people who will use it, and a handover that makes the next iteration manageable. This approach keeps the final result grounded in what the team can verify and continue using.
I write landing-page copy and plan search campaigns for independent opticians whose online enquiries do not explain which appointments they offer.
I design operating models and delivery playbooks when a team needs a practical workflow rather than a vague recommendation. The work includes the implementation itself and concise notes on decisions, so later changes do not rely on one person’s memory. This approach keeps the final result grounded in what the team can verify and continue using. Teams receive the completed work with the key assumptions and testing steps documented, so they can maintain it after delivery.