ML Engineer to create recipe parsing algorithm

Looking for a machine learning engineer to develop a robust model that will: Using language models/NLP/NLU, or related algorithms, and: - Given unstructured recipes extracted as text, or text arrays (e.g. one ingredient per line) - Extract the ingredients, associated quantities, and other elements of the individual recipe / instructions. - Provide the working model wrapped in an API interface that will accept a text blob or array of ingredients and instructions containing the recipe, and return the resultant list of ingredients/quantities, instructions. Said model should be hostable (preferably with an easy install/docker-compose script) on an AWS EC2 or Azure VM (Linux based). Some completion / acceptance requirements: Ingredient parsing must occur at greater than 70% accuracy. API response time (from recipe submitted to response at client) must be < 10s API must be able to handle 100 requests per second (peak). (We will build out/provide the production infrastructure and environment) All documentation, source, models (including installation), and unencumbered intellectual property must be provided. Common cyber security standards must be applied. Any dependencies must not have known vulnerabilities (e.g. CVEs). We will provide: 1) A training/test corpus of recipes 2) A list of ingredients/quantities from our existing ingredients set. 3) Training data in the form of already extracted ingredients/quantities from recipes (recipes and results) that were performed manually (about 2000 recipes). Documentation on the implementation/installation of the environment, and all original models, work products and everything we need to continue on this effort will be required. In your proposal - please indicate timeline, costs, requirements from us, delivery format, your background/portfolio, and any thoughts on the feasibility of this project. Proposals which do not include clear evidence that the bidder has read these requirements will be immediately discarded. Apologies. Please bid the actual amount you are willing to provide this project for. [Note] Must contain job history portfolio or url similar to this.