I will evaluate your AI model and optimize prompts

If you’re building an AI product, chatbot, or workflow, getting a response that “sounds right” isn’t enough. The real question is whether the model consistently follows your instructions across different inputs. That’s where I can help. I work with AI startups, developers, researchers, small businesses, and teams that need their AI outputs tested, evaluated, and improved before they reach users. I use ChatGPT and Claude to test prompts across varied inputs rather than judging a model from one successful response. I look for hallucinations, missed instructions, inconsistent answers, incorrect formatting, weak reasoning, and recurring failure patterns. When I find a problem, I trace it back to the prompt or workflow, test a revised version, and compare the results. The goal is not simply to make one answer better, but to make the overall system more reliable and consistent. I have applied this process to a real hackathon workflow. I tested 30+ AI outputs using different inputs, identified recurring errors, and refined the prompts based on what the testing revealed. The improved approach reduced errors by 64% compared with the previous year. I also refined the prompts further so the workflow could continue improving for future hackathons, even after I was no longer involved in organizing them. Depending on your project, you can receive an evaluation of your AI outputs, tested prompt variations, identified failure patterns, before-and-after comparisons, and practical recommendations for improving reliability. I can also help structure prompts so they are clearer, more specific, and easier to test. Whether you need a chatbot evaluated, an existing prompt improved, or an AI workflow stress-tested with different inputs, I focus on measurable improvements rather than simply making AI-generated text sound better.