@server_kota Profile picture

Kota K.

@server_kota

Developer: worked in startups/big tech. Author of: https://t.co/iGiUvxH7DQ. ⚡ Tips on AWS cloud/SaaS/development.

Similar User
addamsson photo

@addamsson

Arnau Orriols photo

@Arnau_Orriols

Nathan Behary @Nbehary@techhub.social photo

@NathanBehary

David De Miguel photo

@jamesbluecrow

Pinned

Tech stack of a simple SaaS for AWS Cloud Wrote a blog post about it Hope somebody will find it useful ;) saasconstruct.com/blog/the-tech-… #aws #saas #startup #indiechakers #buildinpublic

Tweet Image 1

The OpenAI assistant API (currently in beta) is really good. Cool stuff for GenAI applications. Out-of-the-box: - Serverless vector database (file search). Just upload documents and chat with them. - Code interpreter - Function calls. Assistant can make calls to your backend.…


easy way to have (almost) free vector database for your AI applications: - get LanceDB (open source) - put it in dockerized AWS lambda (1 free mln calls per month) - put DB binaries to s3 (just db.connect("s3...")) that's it #rag #aws


awesome news DynamoDB is going to be cheaper Effective 1st of November 2024- 50% reduction on On Demand throughout and up to 67% off Global Tables. source: aws.amazon.com/about-aws/what… #aws


if you do proxy lambda integrations with AWS ApiGateway, you will have to handle CORS yourself on your backend, meaning you need to include the response header Access-Control-Allow-Origin. #aws


added the product hunt link to the website :) #buildinpublic

Tweet Image 1

There is a solution to AWS Bedrock quota problem for LLM models I mentioned a couple of days ago (a lot of accounts suddenly had their quotas set to 0 in most regions). Tested this solution, it does work. - Go to support in your AWS account - Create a case (choose billing…


According to /aws subreddit and recent QA posts, it seems tons of AWS accounts quotas for AWS Bedrock (LLM models), including old accounts with billing history, were set to 0 in the last 2 weeks. repost.aws/questions/QUK8… #aws


if you need a serverless vector database for your GenAI app, you can host @lancedb in dockerized AWS Lambda. It will scale down to zero when not used. #aws


AWS API gateway has a timeout limit of 29 seconds. Hm... Maybe will try Lambda Function URL... #aws


If you are ever bored with web dev, switch to ML or data engineering, it is so much fun! #buildinpublic


Here are the tips to avoid surprise bills on AWS: - billing alarms - traffic alarms for your website and your backend public endpoint - rate limits on your public backend endpoint - CORS - CDN for your website - cache for your backend #aws #buildinpublic


AWS makes it really easy to build something for fun without costing anything. #aws


Thinking of using dockerized AWS Lambda (with SnapStart to avoid cold starts) and S3 to host a vector database. #aws #saas


United States Trends
Loading...

Something went wrong.


Something went wrong.