Ongoing Maintenance & Support
I provide ongoing maintenance and support for production AI systems. Deploying an agent or pipeline is the beginning, not the end. Models change, APIs update, data drifts, and business requirements evolve. I keep your systems running well and adapting to what comes next.
Production systems need ongoing care
AI systems are not "set and forget" — they require active stewardship to maintain quality and reliability.
The AI landscape moves fast. Model providers release new versions, deprecate old ones, and change pricing structures. The data your agents process shifts in distribution over time. Upstream APIs change their schemas or rate limits. Prompts that worked well six months ago may produce worse results today because the underlying model has been updated. Without someone watching and adjusting, performance degrades gradually until the system is no longer fit for purpose.
I offer maintenance engagements that keep your AI infrastructure healthy after the initial build. This isn't passive monitoring — it's active stewardship that includes regular reviews, proactive improvements, and rapid response when something goes wrong. Think of it as having a dedicated AI engineer on call without the overhead of a full-time hire.
What I provide
Comprehensive support that keeps your AI systems performing at their best as conditions change.
Monitoring and alerting
Continuous monitoring of system health, agent performance metrics, error rates, and resource usage. I set up alerts for anomalies and degradation patterns so issues are caught early — ideally before they impact your operations or your users notice anything wrong.
Prompt and logic refinement
Regular review and optimisation of prompts, decision logic, and agent behaviour based on real-world performance data. As edge cases emerge and requirements shift, I update the system's instructions and workflows to handle new scenarios correctly.
Model updates and migration
When model providers release new versions or deprecate existing ones, I handle the transition. This includes evaluating new models against your specific use cases, updating prompts for compatibility, running regression tests, and managing the cutover with minimal disruption.
Performance optimisation
Ongoing work to reduce latency, lower costs, and improve output quality. I analyse usage patterns, identify bottlenecks, optimise caching strategies, and adjust model selection — sometimes a faster, cheaper model handles a task just as well as a larger one.
Incident response
When something breaks, I diagnose and fix it quickly. I provide defined response time commitments based on severity level, clear communication during incidents, root cause analysis afterward, and preventive measures to stop the same issue from recurring.
System evolution
As your business needs change, your AI systems should change with them. I handle feature additions, scope expansions, new integrations, and architectural adjustments so your investment continues to deliver value as your organisation grows and your requirements mature.
Hiring me for Ongoing Maintenance & Support
I keep production agents, pipelines and trained models working as models, APIs and your data change, including retraining a fine-tune when it drifts.
Who it's for
- Teams with an AI system in production and no one to look after it
- Systems I built that need ongoing care
- Fine-tuned models that need retraining as vocabulary or data change
- Teams who want model upgrades evaluated before they're switched on
What you get
- Monitoring and alerting
- Prompt and logic refinement
- Model updates and migrations, checked against evals
- Incident response
- Small enhancements each month
What I need from you
- Access to the system, logs and hosting
- Documentation, if I didn't build it
- A contact for priorities and incidents
- Notice of upcoming changes on your side
How it runs
Rolling monthly, 3-month minimum.
- Month 1
Onboarding
Review the system, set up access and monitoring.
- Monthly
Run
Monitoring, fixes and agreed enhancements.
- Quarterly
Review
Regular review of performance and priorities.
A monthly retainer sized to the system, 3-month minimum, billed in advance.