AI Evaluation for Impact Masterclass with Calibrate | The Delta

We were invited by The Delta (https://thedelta.org.in) team to give a masterclass on evaluating AI systems within the social sector for their Pragati AI cohort (https://thedelta.org.in/pragati-ai-fo...) and we also presented how Calibrate helps operationalise the best practices in AI evals. We also showed how a non-technical domain expert can run the evals purely by talking to their preferred AI tool (e.g. Claude) using our Skills and CLI. Slides: https://www.amandalmia.com/Nudge-talk/ Calibrate is an open-source tool for evaluating AI systems that we built at ARTPARK and is being used by multiple nonprofits, across India and Africa, to ensure their AI products work reliably. App: https://calibrate.artpark.ai Documentation: https://docs.calibrate.artpark.ai/ Agent Skills: https://docs.calibrate.artpark.ai/age... Learning resources: https://calibrate.artpark.ai/learn Other links from the talk: FormBharo, platform for building form-filling voice agents: https://formbharo.artpark.ai Benchmark for evaluating the voice agent: https://formbharo.artpark.ai/dataset About the speakers: Aman: Website: https://amandalmia.com LinkedIn:   / aman-dalmia   Jigar: Website: https://www.jigarkdoshi.com LinkedIn:   / jigarkdoshi   WhatsApp community on AI evals: https://chat.whatsapp.com/JygDNcZ943a...