paul_simmering Profile Banner
Paul Simmering Profile
Paul Simmering

@paul_simmering

Followers
123
Following
1K
Media
2
Statuses
89

Senior Data Scientist at @Q_InsightAgency Developer of textunnel, a Python package for efficient use of the OpenAI API Interests: MLOps, NLP | Databases

Ludwigsburg, Baden-Württemberg
Joined December 2014
Don't wanna be here? Send us removal request.
@paul_simmering
Paul Simmering
1 year
Quarto is one of the most underrated tools in data science. It’s an all-in-one technical publishing system that lets you weave together code, graphs and markdown text. I use it for everything:. - Reports (PDF).- Dashboards.- Notebooks.- Papers.- My website.
0
0
1
@paul_simmering
Paul Simmering
1 year
There are 10 Python libraries for structured LLM output. I've tried them all. My top recommendations: outlines and instructor. Read my full review here:.
1
5
14
@grok
Grok
2 days
Join millions who have switched to Grok.
114
185
2K
@paul_simmering
Paul Simmering
2 years
Fantastic updates from OpenAI! I've already opened 5 Github issues in my package texttunnel to integrate the improvements. Makes batched NLP even faster. - new models: faster, longer context.- json mode for guaranteed parseable json.- seed for reproduction.
0
0
1
@paul_simmering
Paul Simmering
2 years
RT @fchollet: To build tall, you first need solid foundations. There's no bigger productivity hack than creating components you can keep bu….
0
70
0
@paul_simmering
Paul Simmering
2 years
Important milestone for AI in Germany: a high quality pretrained open LLM 🇩🇪.Super impressed by @laion_ai.
@laion_ai
LAION
2 years
We release LeoLM, a series of German speaking chat models based on Lama 2 7b and 13b. They preserve their ability to understand english text. We also also versions that can reply in English and in German & perform translation between these languages:
0
0
0
@paul_simmering
Paul Simmering
2 years
How to use GPT-4 efficiently in markt research? Have a look at our introduction of texttunnel, featuring a case study from pharma web research.
0
0
0
@paul_simmering
Paul Simmering
2 years
Announcing texttunnel, a package for efficient use of the ChatCompletion API. Did you ever run into these problems?.- slow, sequential requests.- rate limit errors.- paying for the same request twice.- invalid response formats.Then have a look!.
github.com
Python package for extractive NLP using the OpenAI API - qagentur/texttunnel
0
0
0
@paul_simmering
Paul Simmering
2 years
RT @Q_InsightAgency: We are hiring. (Senior) Research & Insight Manager (w/m/d) mit Fokus auf qualitative Forschung .
0
1
0
@paul_simmering
Paul Simmering
2 years
RT @Q_InsightAgency: Did you know that #TikTok posts get 5x more engagement than Instagram posts on average?😮 .Our latest free market resea….
0
2
0
@paul_simmering
Paul Simmering
2 years
Hey @raycastapp I tested the Raycast AI feature in a data science workflow and wrote about it on my blog: I liked it!.
0
0
1
@paul_simmering
Paul Simmering
2 years
How useful are AI assistants like GPT-4 and Github Copilot in a real data science workflow?.I tested them on 5 real tasks and point out the strengths and weaknesses. Read on for my thoughts on what sets successful data scientists apart in an AI-world.
0
0
1
@paul_simmering
Paul Simmering
3 years
@bernhardsson Thanks for the onboarding call in September! Using Modal is a blast.
0
0
1
@paul_simmering
Paul Simmering
3 years
Check out my latest blog post on how to set up a data collector for the Twitter API. It collects up to 500k tweets per month and costs less than $1 to run. Plus, thanks to @modal_labs it's just 100 lines of Python to deploy and monitor in the cloud.
1
1
0
@paul_simmering
Paul Simmering
3 years
RT @Q_InsightAgency: In this blog article we demonstrate the use of text mining as an online research method for gaining #insights into the….
0
1
0
@paul_simmering
Paul Simmering
3 years
Investing time into learning data science skills can feel like a gamble. Will I get to use tool X in the long run? Will model Y stay relevant?.My latest blog post offers a decision framework and my top picks of golden long-term skills
0
1
0
@paul_simmering
Paul Simmering
3 years
LLMs are expensive to run. But their outputs can be used to train smaller, more efficient models. Here, I've asked #ChatGPT to respond like an aspect-based sentiment analysis model. The examples are from Yang and Li's 2022 paper on ABSA.
Tweet media one
1
1
2
@paul_simmering
Paul Simmering
3 years
Thank you @humanloopml for letting me test Programmatic! It's cool to see a glimpse of next generation labeling for NLP.
0
0
4
@paul_simmering
Paul Simmering
5 years
The echarts4r package by @jdatap is a gift that keeps giving! To learn how to use the callbacks feature, I built a simple Shiny app: Callbacks are a great way to connect echarts with other Shiny elements, for example DT tables. #rstats.
0
2
15
@paul_simmering
Paul Simmering
9 years
@pgbovine listening to podcasts on iPhone with Overcast while taking walks, doing chores, on train and bus and while playing videogames.
0
0
0
@paul_simmering
Paul Simmering
9 years
@pgbovine try dropping the 5 to 10 worst spammers :D.
0
0
0