Project 1 Sprint 4

This agile sprint is a continuation sprint3 found here, which itself was a continuation of sprint2 found here and  sprint1 found here. An instructor's solution for sprint2 will be published shortly after this assignment, at which time I will be unable to accept any late sprint2 submissions.

Summary: In this final sprint we will put it all together and create a program which will let a user selected a job and use an AI/LLM to help write a targeted resume/cover letter.

Due: Sunday March 9th at 11:59pm (Spring Break officially begins on Monday March 10th)

Important: Do all of your work in the same github project as sprint1 Sprint2 and Sprint3.

Deliverable for Sprint4

• In this final Sprint you will build a complete application which will
• Allow the user to select a job and display the full job information as per sprint3
• Allow the user to enter user data as per sprint3
• keep track of the user data using a 'profile name' (no need for security - we'll assume only one user for now)
• allow the user to enter multiple profiles with distinct data - the Database you vs the Full-Stack-Developer you for example
• Allow the user to select one of these users/profiles from a list
• when the user has selected one job and one user profile, Programmatically send the job description and the information about the user (including projects, classes etc) to the LLM/AI that you worked with in sprint1 (or another one if you have changed for some reason - please inform your instructor in this case)
• Ask the LLM to create a customized cover letter in markdown format for this user and this job, save the resulting markdown document to a file (only the file, and not any of the extra stuff that the AI system sends back)
• Ask the LLM to then create a customized resume in markdown format for this user and this job. save the resulting markdown document to a file (only the file, and not any of the extra stuff that the AI system sends back)
• programmatically convert the markdown to PDF format
• for example in go one might use:
• https://github.com/mandolyte/mdtopdf
• or
• https://github.com/stephenafamo/goldmark-pdf
• in python one might use:
• https://pypi.org/project/markdown-pdf/
• or the slightly less well documented
• https://github.com/jmaupetit/md2pdf
• Automated Tests
• test that you get a 200/ok back from the LLM when you query it (you will need to use the github secrets approach for this)
• test that your automatically created prompt contains both the job description and then user information
• this will require either multiple tests or a very long one.
• write at least one other test of one of your production functions
• Readme:
• update your readme to explain which of your tests functions cover which of the case  in the test section above.
• add instruction so I know everything I need to in order to run your program properly and painlessly on my machine.
• Evaluation (DUE in class the second class of the week after spring break break as a physical paper):
• Write a paper evaluating your experiences with the LLMs.
• How well did the LLM work
• did it provide a reasonable cover letter/resume given the information about you that you gave it and the job description
• Did it add anything else irrelevant (for example in one of my test runs it added that I was proficient in Spanish - I am not and did not tell the LLM that I was)
• Evaluate these LLMs for the potential to change the industry
• Sequoia Capital has found that there has been $600 billion invested globally in AI, but less than $100 billion in revenue (not profit, but revenue) from AI LLMS, that means either we are in a big bubble, or we need to find some money making applications for it; argue whether we are at the cusp of something new or in a bubble based on your experiences in this assignment and any previous experience (not just previous experience)
• in the first class of the week after spring break, we will do our standard retrospective. In that retrospective, in addition to the usual things, I was to have you discuss your results with each other. How similar were the results? Do you think you would be able to detect AI generated cover letters/resumes if you were a recruiter getting a hundred of each for a job? Discuss this final question in your paper.

Submitting this assignment

At the due date (or since it is midnight - probably the next morning) I will re-pull your project from the original repository for the readme and program part

I will take the written papers (in standard Business English) in the second class in the week after spring break.