Context
Key Performance Indicators :
Understanding Prompt ranking
Understanding prompt matching
Ability to reuse previous knowledge
Wednesday
RAG Evaluation Data Generation (Rehmet)
Understanding of prompt matching and ranking (Emtinan)
Thursday
RAG evaluation metrics (Emtinan)
RAGOps - DevOps of RAG development and production deployment (Rehmet)
DELIVERABLES
NOTE : Document should be a PDF stored in google drive or published blog link . DO NOT
SUBMIT A LINK as PDF! If you want to submit pdf document , it should be the content of
your report not a link .
Interim Submission - Wednesday 8pm UTC
Link to your code in GitHub
Repository where you will be using to complete the tasks in this week ' s
challenge . A minimum requirement is that you have a well structured
repository and some coding progress is made .
A review report of your reading and understanding of Task 1 and any progress you
made in other tasks .
Feedback
You may not receive detailed comments on your interim submission but will receive a
grade .
Final Submission - Saturday 8pm UTC
Link to your code in GitHub
Complete work for Automatic prompt generation
Complete work for Automatic evaluation
Complete work for Evaluation Data Generation
6/6/24, 11:51 PM 10 Academy Cohort B Weekly Challenge: Week 7
https://tenx.10academy.org/trainee/challenge/10 Academy Cohort B Weekly Challenge: Week 7-lwythrp7---75 9/1010 Academy Cohort B Weekly
Challenge : Week 7
Precision RAG : Prompt Tuning For Building
Enterprise Grade RAG Systems
BUSINESS OBJECTIVE
PromptlyTech is an innovative e - business specializing in providing AI - driven solutions
for optimizing the use of Language Models (LLMs) in various industries . The company
aims to revolutionize how businesses interact with LLMs , making the technology more
accessible , efficient , and effective . By addressing the challenges of prompt engineering ,
the company plays a pivotal role in enhancing decision - making , operational efficiency ,
and customer experience across various industries . PromptlyTech ' s solutions are
designed to cater to the evolving needs of a digitally - driven business landscape , where
speed and accuracy are key to staying competitive .
The company focuses on key services : Automatic Prompt Generation , Automatic
Evaluation Data Generation , and Prompt Testing and Ranking .
1 . Automatic Prompt Generation Service :
This service streamlines the process of creating effective prompts , enabling
businesses to efficiently utilize LLMs for generating high - quality , relevant content .
It significantly reduces the time and expertise required in crafting prompts
manually .
2 . Automatic Evaluation Data Generation Service :
PromptlyTech ’ s service automates the generation of diverse test cases , ensuring
comprehensive coverage and identifying potential issues . This enhances the
reliability and performance of LLM applications , saving significant time in the
QA(Quality Assurance) process .
3 . Prompt Testing and Ranking Service :
6/6/24, 11:51 PM 10 Academy Cohort B Weekly Challenge: Week 7
https://tenx.10academy.org/trainee/challenge/10 Academy Cohort B Weekly Challenge: Week 7-lwythrp7---75 1/10Semantic Similarity Matching: Using NLP techniques to match prompts based on
semantic content , ideal for understanding nuanced differences in prompt
effectiveness .
You should adopt an innovative approach to prompt evaluation by utilizing Monte Carlo
matchmaking and ELO rating systems , or any alternative method to match and rank .
Task 5 : User Interface Development
Develop a user - friendly interface for interacting with the prompt engineering system .
UI Design : Plan and design a user interface that allows users to easily input data ,
receive prompts , and view evaluation results .
UI Implementation : Develop and integrate the user interface with the backend
prompt engineering system .
Task 6 : System Integration and Testing
Integrate all components of the system and conduct comprehensive testing .
Integrate the prompt generation , Evaluation Data Generation , evaluation , and user
interface components .
Test the entire system for functionality , usability , and performance . Refine based
on feedback and test results .
TUTORIALS SCHEDULE
In the following , the Bold indicates morning sessions , and Italic indicates afternoon
sessions .
Monday : Understanding Prompt engineering
Here the trainees will understand the week ’ s challenge .
Introduction to Week Challenge (Yabebal)
Introduction and challenges to prompt engineering (Emtinan)
Key Performance Indicators :
Understanding week ’ s challenge
Understanding the prompt engineering
Ability to reuse previous knowledge
Tuesday
RAG components (Rehmet)
Techniques to improve R (Retrievers) in RAG (Emtinan)
6/6/24, 11:51 PM 10 Academy Cohort B Weekly Challenge: Week 7
https://tenx.10academy.org/trainee/challenge/10 Academy Cohort B Weekly Challenge: Week 7-lwythrp7---75 8/10when to use it
Community prompt sourced from the open-source GitHub repo GetachewAbebe/Prompt_Engineering_using_Precision_RAG (MIT). A "Context" style prompt — adapt the placeholders and specifics to your task. Imported as-is and not independently retested here, so check the output before relying on it.
tags
codingcommunitydeveloper
source
GetachewAbebe/Prompt_Engineering_using_Precision_RAG · MIT
more in Coding
Coding✓ tested
Senior code review (strict mode)
senior staff engineer running a merciless but fair review
Coding✓ tested
Debug by hypothesis, not by guessing
debugging partner who forms theories before touching code
Coding✓ tested
Generate tests from described behavior
test engineer who writes tests that would actually catch regressions