Upwork is hiring a NLP text analysis and prompt recommendations

NLP text analysis and prompt recommendations

Upwork  ·  US  ·  $37k/yr - $83k/yr
over 1 year ago

Asignment.

Looking for an NLP specialist or data scientist to analyze user input for guided meditations co-writer in order to:

- identify parameters like theme, phrases, tone of voice, style,etc

- based on key parameters make specific adjustments to prompt

Context:

There is a GPT meditation script co-writer where user generates paragraphs of guided meditation for himself and Chat GPT follow-up with its paragraphs.

Scope:

Identify Common Themes and Keywords: Analyze the content to pinpoint recurring themes (e.g., nature, mindfulness, self-reflection) and frequently used keywords that signify user preferences.

Determine Tone and Style: Assess the tone (e.g., calming, uplifting) and style (e.g., narrative, instructive) prevalent in the user submissions. This includes examining the language complexity, sentence structure, and any specific linguistic patterns.

Extract User-Specific Preferences: Look for any personalized elements or unique preferences indicated by users, such as preferred meditation duration, specific goals (e.g., stress reduction, focus enhancement), or favored imagery (e.g., ocean waves, forest walks).

Recommendations for Prompt Adjustments: Based on the analysis, suggest specific adjustments to the prompts to use with ChatGPT. These recommendations might include:

Themes to Emphasize: Suggesting to incorporate certain themes more heavily or introducing new ones that align with user interests.

Tone and Style Guidelines: Advising on the desired tone (e.g., more soothing or affirmative) and style (e.g., using more descriptive language or simple, direct instructions) to match user preferences.

Personalization Elements: Proposing the inclusion of elements like user-specified goals, preferred imagery, or other personalized details to make the meditations feel more tailored.

Job is closed

This job is already closed and no longer accepting applicants, sorry.