AI Backend Integration with SQL, AWS Cognito & S3
Budget: €1,500 – €3,000 EUR
Hi, I’m Josef from CEO of BioSonic
Here is a brief of our project. I also attached the rough architecture.
We want to get started right away and are actively interviewing candidates
Let me know if you have any other questions!
We've built a portal frontend launched on Vercel and now want to take in users who login through our AWS Cognito, upload .wav files to our S3 bucket where our AI model is running inference (already implemented) and then build the backend that handles diffrent users, diffrent outputs from AI inference and serves this to the frontend
The frontend is already in React, so you’ll implement pre-signed URL uploads to S3, handle the flow from AWS Cognito validation, and build a Supabase/PostgreSQL API to store customerUUID, rowID, file paths, spectrogram paths, species and timestamps.
Our AI inference pipeline already processes .wav files and saves results in S3 and Supabase. You’ll integrate AWS Cognito for authentication and create the endpoints getPresignedURL, getCustomerRows, getCustomerRowDetails, updateCustomerSpeciesIdentification, getAudio and getSpectrogram. We’re looking for someone with 5+ years’ experience in S3, Supabase/PostgreSQL, API development and AWS Cognito, with clear API design. Serverless (Lambda, Vercel) and spectrogram/FFT experience are a plus. The project starts immediately and runs two weeks: by May 7 five users should be able to upload and view results, and by May 14 the full solution must be live.
Specialization
Back-end Web Developer
Experience level
Senior
Core skills:
JavaScript
React
Next.js
Amazon Cognito
Amazon S3
Supabase
API
AWS
PostgreSQL
Nice-to-have skills:
AWS Lambda
Serverless Computing
Industry experience
Some Q/A about the project follows
Q: Is FFT/Spectrogram just a file stored in S3, or is any image preview/rendering expected on frontend?
A: Right now the MVP is just a png of the spectogram, but our prio 2 (after this project) is to render spectograms
Q: Do you have the Supabase tables already set up, or should I provision the schema as part of the work?
A This is not set up yet and it's part of the work
Q: For S3: Will I have IAM access to configure and test the presigned URL flows?
A: Yes
Here is a brief of our project. I also attached the rough architecture.
We want to get started right away and are actively interviewing candidates
Let me know if you have any other questions!
We've built a portal frontend launched on Vercel and now want to take in users who login through our AWS Cognito, upload .wav files to our S3 bucket where our AI model is running inference (already implemented) and then build the backend that handles diffrent users, diffrent outputs from AI inference and serves this to the frontend
The frontend is already in React, so you’ll implement pre-signed URL uploads to S3, handle the flow from AWS Cognito validation, and build a Supabase/PostgreSQL API to store customerUUID, rowID, file paths, spectrogram paths, species and timestamps.
Our AI inference pipeline already processes .wav files and saves results in S3 and Supabase. You’ll integrate AWS Cognito for authentication and create the endpoints getPresignedURL, getCustomerRows, getCustomerRowDetails, updateCustomerSpeciesIdentification, getAudio and getSpectrogram. We’re looking for someone with 5+ years’ experience in S3, Supabase/PostgreSQL, API development and AWS Cognito, with clear API design. Serverless (Lambda, Vercel) and spectrogram/FFT experience are a plus. The project starts immediately and runs two weeks: by May 7 five users should be able to upload and view results, and by May 14 the full solution must be live.
Specialization
Back-end Web Developer
Experience level
Senior
Core skills:
JavaScript
React
Next.js
Amazon Cognito
Amazon S3
Supabase
API
AWS
PostgreSQL
Nice-to-have skills:
AWS Lambda
Serverless Computing
Industry experience
Some Q/A about the project follows
Q: Is FFT/Spectrogram just a file stored in S3, or is any image preview/rendering expected on frontend?
A: Right now the MVP is just a png of the spectogram, but our prio 2 (after this project) is to render spectograms
Q: Do you have the Supabase tables already set up, or should I provision the schema as part of the work?
A This is not set up yet and it's part of the work
Q: For S3: Will I have IAM access to configure and test the presigned URL flows?
A: Yes