Develop software code leveraging AI, ML and NLP to identify headcount and non-headcount reduction opportunities at a Company based on its financial profile
Budget: $30 – $250 USD
Use Case: Develop software code leveraging AI, ML and NLP to identify headcount and non-headcount reduction opportunities at a Company based on its financial profile
Proposed Cost: $150 - $200
Input:
Income Statement (either xls or pdf annual report)
This statement usually has 4 or 5 cost categories such as Technology, Operations, G&A, etc.
Benchmarking database (xls document which describes income statement of companies in various industries)
I will provide a benchmark database and provide a criteria for selecting the benchmark to compare
UI input to provide specific parameters
Examples of parameters include “Industry”, “Sub-Industry”, etc.
Annual Report or Investor memorandum (pdf document)
This is a 50 - 60 page pdf document describing the company and its performance in significant details.
Output:
A table structure that shows areas where company’s cost is higher than benchmark companies
I will provide a criteria for identifying areas where company’s cost are higher than benchmark
Commentary on (a) potential reasons for higher costs in certain areas; (b) ways to bring down the cost
I will provide a database of “reasons costs are higher than benchmark” and a database of “ways to bring down cost”
Commentary summarizing company profile using BERTA, GPT-3 or similar NLP models
Plan is to fine-tune and train the models in subsequent phases but not in this phase
Deliverables: Provide a fully functional application that takes input and provides output as described above:
Develop a basic UI that can be used to upload the Input described above
Develop python code that
Creates the output as described above
Leverages NLP models such as BERT, GPT-3,RoBERTa T5 or XLNet to summarize information
Uses OpenAI API or the Hugging Face's Transformers library
Note - My plan is to - at a later stage, not in this stage - preprocess and prepare dataset for fine-tuning the model
Develop API using frameworks such as Flask or FastAPI and deploy API using AWS Lambda or Google
Proposed Cost: $150 - $200
Input:
Income Statement (either xls or pdf annual report)
This statement usually has 4 or 5 cost categories such as Technology, Operations, G&A, etc.
Benchmarking database (xls document which describes income statement of companies in various industries)
I will provide a benchmark database and provide a criteria for selecting the benchmark to compare
UI input to provide specific parameters
Examples of parameters include “Industry”, “Sub-Industry”, etc.
Annual Report or Investor memorandum (pdf document)
This is a 50 - 60 page pdf document describing the company and its performance in significant details.
Output:
A table structure that shows areas where company’s cost is higher than benchmark companies
I will provide a criteria for identifying areas where company’s cost are higher than benchmark
Commentary on (a) potential reasons for higher costs in certain areas; (b) ways to bring down the cost
I will provide a database of “reasons costs are higher than benchmark” and a database of “ways to bring down cost”
Commentary summarizing company profile using BERTA, GPT-3 or similar NLP models
Plan is to fine-tune and train the models in subsequent phases but not in this phase
Deliverables: Provide a fully functional application that takes input and provides output as described above:
Develop a basic UI that can be used to upload the Input described above
Develop python code that
Creates the output as described above
Leverages NLP models such as BERT, GPT-3,RoBERTa T5 or XLNet to summarize information
Uses OpenAI API or the Hugging Face's Transformers library
Note - My plan is to - at a later stage, not in this stage - preprocess and prepare dataset for fine-tuning the model
Develop API using frameworks such as Flask or FastAPI and deploy API using AWS Lambda or Google