HyperProphet – The Easiest Way Run FB Prophet at Scale
We are Anand and Raghav, the creators of HyperProphet (https://hyperprophet.com/), a tool to enable developers to compute thousands of forecasts in minutes using fbprophet without worrying about infrastructure. Scaling forecast computation is a hard problem, often costs developers in productivity. Even established organizations spend a lot of time trying address this, but often the outcomes are not easy to use and manage[1][2]. We spent the last few years building a demand planning product for CPG and retail in which we had to forecast thousands of time-series. Computing such…
What it does
In the maker’s words, at launch
We are Anand and Raghav, the creators of HyperProphet (https://hyperprophet.com/), a tool to enable developers to compute thousands of forecasts in minutes using fbprophet without worrying about infrastructure. Scaling forecast computation is a hard problem, often costs developers in productivity. Even established organizations spend a lot of time trying address this, but often the outcomes are not easy to use and manage[1][2]. We spent the last few years building a demand planning product for CPG and retail in which we had to forecast thousands of time-series. Computing such huge number forecasts in a reasonable time was a challenge. Facebook’s Prophet was one of our work-horse algorithms. Although Prophet made many things easy, it was still slow for a large dataset. Speed was critical for quick experimentation. To improve the speed, we built a small library to run compute on AWS Lambda. Encouraged by the early results, we built an in-house service HyperProphet which we are sharing with you today. The key features of HyperProphet are: - Fast: compute thousands of forecasts in minutes - Simple: switch from Prophet to HyperProphet by changing a couple of lines of code - Productive: experiment with forecasts in your notebooks without worrying about infrastructure HyperProphet is in early beta, and we are looking for more people to try it and give us feedback. We would love to see it used in diverse use cases and see how it performs. We see possible use-cases across multiple industries for capacity planning, anomaly detection, financial planning, among others. If you have any thoughts, we’d love to hear them in the comments! [1]: https://www.slideshare.net/MahanHosseinzadeh/prophet-at-scale-using-prophet-at-scale-to-tune-and-forecast-time-series-at-spotify [2]: https://medium.com/walmartglobaltech/scaling-machine-learning-algorithms-fbprophet-xgboost-with-pyspark-on-w-mlp-405fadca1c19
Does the same job
all alternatives →
- PIPortfolioPilot – intelligent portfolio management for every investor2022 · ▲7
Hey HN! I'm Alex, one of the co-founders at Global Predictions. We are officially launching PortfolioPilot today! Since starting the company 2 years ago, we’ve been working tirelessly to build a personal portfolio management platform that empowers everyday people to feel more confident investing. We focused on aggregating your entire net worth, evaluating across a set of standardized metrics, and offering suggestions based on our commercial-grade Macro Insights & Recommendation Engine. We started by building a product we wanted to use and iterated and refined based on early user feedback. We…
- S1Serve 100 Large AI models on a single GPU with low impact to TTFTNov 2025 · github.com · ▲7
I wanted to build an inference provider for proprietary AI models, but I did not have a huge GPU farm. I started experimenting with Serverless AI inference, but found out that coldstarts were huge. I went deep into the research and put together an engine that loads large models from SSD to VRAM up to ten times faster than alternatives. It works with vLLM, and transformers, and more coming soon. With this project you can hot-swap entire large models (32B) on demand. Its great for: Serverless AI Inference Robotics On Prem deployments Local Agents And Its open source. Let me know if anyone…
More work this month
the category →
The app store for voice native apps that lives in your notch
Work · 28d ago · voiceos.com
The New Calendly▲211Handle all of the work before, during, and after meetings
Work · 17d ago · calendly.com
Launched alongside, August 2020
the whole month →
- IB
Life & fun · 2020 · storycreatorapp.com
