Amanda JohnsonMBA, MSM, PMP, PSM I, LSSYB| NoneMontevideo, Departemento de Monevideo, Uruguay
I'm hoping I can get hired by a company that is starting to embrace GenAI. Saving Changes...
Elizabeth MassuraPrincipal Marketo Consultant| AcxiomChicago, IL, United States
I haven't yet been using Gen AI tools on a regular basis, but it's been interesting to see how Gen AI has been incorporated into tools I already use. For example, Domo has Gen AI to develop calculated fields for use in data visualization. Last week I attended an event with Salesforce on using their standard AI agents and creating prompt templates to leverage Gen AI for personalized emails. Saving Changes...
Patricia WhiteEducator/Trainer| UMUCOrange Park, Fl, United States
I think all AI tools are indispensable because they work and some work better than others. I like Chat GPT because you learn so much and it assists in solving real problems. Saving Changes...
I’m glad to share a range of indispensable tools and resources for working with Generative AI data. Here’s an overview of the tech stack and processes that can boost your AI projects:
1. Data Collection & Synthetic Data Tools
Purchasing Synthetic Data: For privacy or scarcity issues, platforms like Synthesis AI and Datagen provide high-quality synthetic datasets for training models, especially in computer vision. Mostly AI and Tonic AI are also useful for generating structured synthetic data.
Open Datasets: Sites like Kaggle and UCI Machine Learning Repository offer a rich repository of real-world datasets for model training and experimentation.
2. Data Cleaning and Preprocessing
Python Libraries: Libraries like Pandas and NumPy are staples for data wrangling and cleaning. Dask helps manage larger-than-memory datasets, and PyJanitor provides additional tools for easy data cleaning.
Data Profiling: Pandas Profiling and Great Expectations are great for automating data quality checks and generating data reports.
Data Version Control: DVC (Data Version Control) is a powerful tool for versioning datasets and tracking changes, making it easier to manage evolving data needs.
3. Model Deployment & Fine-Tuning
Model Training & Deployment: For deploying and fine-tuning models, I recommend frameworks like Hugging Face Transformers for NLP tasks and PyTorch or TensorFlow for custom model development. Weights & Biases or MLflow are invaluable for tracking experiments and managing hyperparameters.
APIs & Integration: FastAPI or Flask can be used to build REST APIs for deploying models quickly. AWS SageMaker, Google Vertex AI, and Azure ML are comprehensive platforms that support end-to-end model management.
4. Data Cleaning Processes
Automated Tools: Trifacta and Talend are useful for data wrangling and transformation at scale, especially when dealing with large datasets.
Manual Validation: Despite automation, manual review remains crucial for validating data quality, especially in critical applications like healthcare or finance.
5. Charting & Data Visualization Tools
For Exploratory Data Analysis (EDA): Tools like Matplotlib, Seaborn, and Plotly in Python are staples for interactive and static visualizations. Altair is another excellent library for generating declarative statistical visuals with minimal code.
Dashboards: Tableau, Power BI, and Google Data Studio are perfect for creating shareable dashboards. Streamlit is a fantastic open-source option for building interactive web apps for data science projects quickly.
Geospatial Data: If you’re working with location-based data, Kepler.gl and GeoPandas are invaluable for mapping and geographic analysis.
6. Collaboration & Reproducibility
Jupyter Notebooks: For quick prototyping and sharing insights, Jupyter and JupyterLab are essential. They integrate well with visualization libraries for exploratory work.
Version Control & Collaboration: GitHub or GitLab paired with platforms like Google Colab or Kaggle Notebooks makes it easy to collaborate and share experiments.
Recommendations for Enhancing Gen AI Capabilities
Experiment Management: Incorporate tools like Weights & Biases to monitor performance metrics and optimize your models efficiently.
Pipeline Automation: Using Apache Airflow or Prefect can help automate your data cleaning and preprocessing workflows, ensuring consistency and reproducibility.
Community Engagement: Engage with communities on Kaggle, Stack Overflow, or Reddit to stay updated on best practices and innovations in Generative AI.
I hope these tools and recommendations help you refine your AI workflows! Saving Changes...
TAOFEEK ADEGBITEProject Engineer/Manager| ULTIMUS CONSTRUCTIONLagos, LA, Nigeria
I make use of chat gpt very for daily site report and template adjustment. Saving Changes...
Salman ChohanSenior Project Manager| TPL MapsIslamabad, IS, Pakistan
Earlier i use chatgpt, but now a days i am using claude.io which i found more professional. And for representation of gadgets in presentation i use miro.com Saving Changes...
I am getting started with AI. I have used ChatGPT and CoPilot. My aim is to build an application using RAG or other pathways. Can someone share their experience and suggest a roadmap I should follow? Also, are there any groups I can join where similar group projects are done? Saving Changes...
In a Microsoft environment, Azure AI and, to some extent, Google Cloud AI have been utilized as cloud platforms.
Power BI works well at providing dashboard views
Saving Changes...
Pham Thi Hai VanHead of Building Service Engineering| Inros-Lackner VietnamHanoi, Viet Nam
Some of the most indispensable tools I use are the large language models from OpenAI, such as GPT-4, and image generation models like DALL·E. These platforms allow me to quickly generate high-quality text and visuals for a variety of purposes—whether it’s automating content creation, generating ideas, or crafting customer-facing communications. Saving Changes...
PMI Team Member
