Ivandros Awom

A short post on the world of tech. If you believe you can simply apply standard software engineering practices to A.I., you have already wasted time and money! If you have leadership/hiring managers interviewing Data Science professionals who has a command on what it truly takes to deploy AI production ready deep learning models for AI application, these hiring manager would not be wasting experienced professionals time asking about their experience with LLM's and OpenAI. Especially when those hiring managers haven't even completed a full cycle of production ready machine learning models that were developed just using a common ML algorithms i.e (Random Forrest). There are true practitioners out here that can lead, let them lead! They are teaching in higher education, and continuing building courses and researching every aspect of the profession and guiding/mentoring a new generation of practitioners. A.I. requires diverse thinking and experience, you must be very resourceful in order to achieve successful A.I. implementation. Every piece has to be assessed and evaluated every step of the way. Invest in your project and team wisely.

10

Here's how you can build boxplots in Python using Seaborn. Boxplots are one of my favorite ways to explore distributions of data, and there are also some very cool variations of boxplots that you might not be aware of. Are you using boxplots in your data analysis? If you want to develop Python for data analysis skills, and are looking for a step by step learning path, check this out... HOW TO LEARN PYTHON: https://lnkd.in/gUNa7H-m #learning #python #data #analyis #datascience #businessintelligence #careers

31

2 Comments

Forget Hugging face ❌ For ML / LLM models !! As a Machine Learning Engineer, managing and loading large language models (LLMs) used to be very tough for me. Recently, I came across @CSGHub—a powerful tool that has made my Workflow very easy: ✅ All your LLM assets—datasets, model files, and code—are neatly managed in one place. ✅ User-friendly interface. ✅ Access CSGHub online, through Git commands, or even via a Chatbot. ✅ Manage your data securely, even when offline. ✅ Connect with a supportive community for ideas and assistance. Why you should switch to CSGHub over Hugging face: ✅ 100% Open Source: Full control with no hidden constraints. ✅ On-Premise Deployment: Keep your data secure within your own infrastructure. ✅ All-In-One Management: Not just LLMs—handle datasets and source code too. If you’re in AI, whether you’re a beginner or an expert, CSGHub will simplify your LLM management and enhance your workflow. Don’t miss out—try it and see the difference. CSGHub GitHub Repo: https://lnkd.in/gNUqUmE7 #CSGHub #AI #MachineLearning #DataScience #DeveloperTools #OpenSource

309

74 Comments

I've found that the most effective way to learn in Data Science is by undertaking a project and doing it through from start to finish. I've created a list of projects/application videos that you can use as a starting point and extend upon to suit your needs. 🌟The Making of SRKGPT: Crafting an AI with Shahrukh Khan's Style LLama v2 Finetuning youtu.be/gYPwx0DR7zc   🌟Building Scalable Query-Item Two-Tower Model based Retrieval System youtu.be/o-pZk5R0TZg 🌟 Enhancing the performance of smaller open source LLM by training on larger LLM dataset like Chat-GPT youtu.be/paGr-t1wSOQ youtu.be/lo11Iczb0Vc  🌟 Finetuning Falcon & LLama-2 LLM’s to custom Datasets  ​​youtu.be/CxqZ5j3xlt0 youtu.be/8cc4bJtycOA 🌟 Leveraging Pre-trained Bert to generate Entity Embeddings across domain: youtu.be/v-0J7o-nDBE 🌟Model Compression with Knowledge Distillation: youtu.be/1N_EBJUOjVU  🌟Dynamic Pricing Strategy to effectively fill slots | E-commerce | Travel | Hospitality Domains: youtu.be/a_CXpnsvPa0 🌟 Train your own Product Embeddings | Embeddings across Domains | Training on 1.02 Billion rows on GPU: youtu.be/DN4S96oHRhE 🌟 Model-Based Account Recommendations @ Twitter | Recommendation System | Retrieval & Ranking | RECSYS: youtu.be/Xqo8fwgjxW4 🌟 How Twitter recommends Tweets: Inside Twitter's Recommendation Algorithm youtu.be/IhGq9jgcxFM 🌟 Advertiser Recommendation Systems at Pinterest | Machine Learning & Algorithm-driven Ads Platform: youtu.be/KEiBJ0fal3Q 🌟 Approximate Nearest Neighbour and a popular Library Annoy for k-Nearest Neighbour Search & Product Quantizer: youtu.be/DSQOrBTqmYA & youtu.be/50PNumB7s3U 🌟 PID Controller to add a layer of diversity to the recommender systems: youtu.be/laTxgnzjfR0 🌟 Train Neural Networks to Approximate any Function youtu.be/4PvGKuqRQTE 🌟 Dimensionality Reduction Techniques: PCA versus AutoEncoders, A comparison through Image Dataset https://lnkd.in/gC6XQfez #datatrek #datascience #machinelearning #statistics #deeplearning #ai

242

6 Comments

Hi there, Are you wondering why there is a need to have Delta Lake , although we have Data lake? If so, follow along i be revealing what i have learned today. Data lake is a storage repository to store huge volumes of raw data. This includes structured,semi structured and unstructured data. On amazon its a S3, Azure its Adls gen2 and on Google its Google Cloud storage (GCS). Overcoming the following Data lake, challenges leads to the adoption of Delta lake. • Acid compliance. • Data quality. • Data consistency. • Data reliability. • No DML operations. Delta Lake is an open source storage layer. It is small utility installed on spark cluster. Delta Lake format is similar to parquet but has it own differences.Below are major differences in both. • It enables Acid compilance. It enables transaction logs. #DeltaLake#Data#DataLake#Fundementals #DataEngineering

5

How I got companies to consider me for a data scientist role without previous experience... ✨My resume strategy and layout was the key! Resume Tips: 📌 Avoid Using a One-Size-Fits-All Resume Customize each resume for the job by using keywords to get past Applicant Tracking Systems (ATS). 📌 Projects First Place projects as the first section of your resume to emphasize them as your main "work experience." 📌 Projects Detail Include at least 3 data science projects with detailed descriptions including technical skills utilized and results with metrics. 📌 Work Experience List your actual work history, highlighting any transferable skills that are relevant to data science. 📌 Education List your educational background and include any related certifications. ✨ This resume layout makes sure your data science projects grab an employer's attention first. P.S. How have you structured your resume to highlight your skills and projects? #DataSistah

67

13 Comments

🔥 𝗪𝗵𝗮𝘁 𝗶𝘀 𝘁𝗵𝗲 𝗱𝗶𝗳𝗳𝗲𝗿𝗲𝗻𝗰𝗲 𝗯𝗲𝘁𝘄𝗲𝗲𝗻 𝗮 𝘀𝗲𝗿𝗶𝗲𝘀 𝗮𝗻𝗱 𝗮 𝗱𝗮𝘁𝗮𝗳𝗿𝗮𝗺𝗲 𝗶𝗻 𝗣𝗮𝗻𝗱𝗮𝘀? 🔹 Dimensions: - Series: One-dimensional labeled array - DataFrame: Two-dimensional labeled structure with different column types 🔹 Functionality: - Series: Limited functionality, mainly for single-column operations - DataFrame: More functionality, like filtering, merging, and joining, perfect for tabular data 🔹 Data Alignment: - Series: Doesn't support alignment - DataFrame: Supports alignment based on column labels 🔹 Performance: - Series: Faster for single-column operations - DataFrame: Slower due to its complex structure 🔹 Index: - Series: Requires an index - DataFrame: Optional index 🔹 Columns: - Series: No columns - DataFrame: One or more columns 🔹 Name: - Series: Optional name for identification - DataFrame: Optional column-wise names 💡 The choice between Series and DataFrame depends on your data analysis needs. Don't be mistaken, they serve different purposes! #DataAnalysis #Python #Panda #MythBuster #Pandas #Series #DataFrame #Functionality #DataAlignment #Performance #Index #Columns #Name #DataAnalysis #Python #Panda #ReachifyMe

2

Dear Aspiring Data Scientist, Explore these high-quality online courses tailored specificaly for you. And guess what? You can enroll for free! Harvard University 1. Statistics and R: https://lnkd.in/ggwKPfH3 2. Probability: https://lnkd.in/gRA-sCrU 3. Linear Regression: https://lnkd.in/gCqYdf_W 4. Inference and Modelling: https://lnkd.in/gSnuYTMW RICE (via Coursera) 1. Introduction to Data Analysis with Excel: https://lnkd.in/gBVHWzvR Johns Hopkins (via Coursera) 1. Advanced Statistics for Data Science Specialization: https://lnkd.in/g_n-3Wn5 Tip: On Coursera, opt for 'Audit the course' to access all learning videos and materials without charge. Hungry for more data analysis courses? You can just peek into my previous posts. Walk with me on this enlightening journey into data analysis right here on LinkedIn. Let's chart this course, learn, and flourish together!

262

10 Comments

Brushing up on some python skills to develop a skills checker. ✅ Have a look at the code here https://lnkd.in/ercsyb3g I want to add this to an interactive website. Does anyone have any experience of Flask and can recommend some helpful resources? 🤔 #ContinuousLearning #Python #CodeFirstGirls #Flask

3

Junior Data Scientist Mindset: "The team needs forecasting so I am going to deliver sophisticated models like LSTM or N-Beats that forecast in real-time! It may take more than 2 months to develop and deploy, but it will be worth it!" Senior/Staff Data Scientist Mindset: "The team currently relies on Excel spreadsheets to calculate moving averages for quarterly forecasting. So, I am going to use a simple ARIMA model to prove the value first. Since they use forecasts quarterly, I can build a batch inference that delivers forecasts quarterly. This will take less than 2 months to develop and deploy." — Ace interviews with these👇 📕 𝗜𝗻𝘁𝗲𝗿𝘃𝗶𝗲𝘄 𝗣𝗿𝗲𝗽 𝗖𝗼𝘂𝗿𝘀𝗲𝘀: https://lnkd.in/gzgB-dHT 📘 𝗝𝗼𝗶𝗻 𝗗𝗦 𝗜𝗻𝘁𝗲𝗿𝘃𝗶𝗲𝘄 𝗕𝗼𝗼𝘁𝗰𝗮𝗺𝗽: https://lnkd.in/eiA5Ntdp 📙 𝗝𝗼𝗶𝗻 𝗠𝗟𝗘 𝗜𝗻𝘁𝗲𝗿𝘃𝗶𝗲𝘄 𝗕𝗼𝗼𝘁𝗰𝗮𝗺𝗽: https://lnkd.in/e6HbN6dy 👉 Smash 👍 and follow Daniel Lee to land dream data & AI jobs 🚀 — Inspired by Benjamin Rogojan Data Engineering post

434

15 Comments

5️⃣ Data Science interview questions to solve this week 𝘛𝘩𝘦𝘴𝘦 𝘯𝘶𝘮𝘣𝘦𝘳𝘴 𝘢𝘳𝘦 𝘮𝘢𝘥𝘦 𝘶𝘱 — 𝘱𝘭𝘦𝘢𝘴𝘦 𝘥𝘰𝘯’𝘵 𝘴𝘶𝘦 𝘮𝘦. ↳ TikTok has noticed a 20% decrease in monthly active users over the last quarter. What do you think could be happening? ↳ During a weekly report, you find that the average session duration on Candy Crush has suddenly increased by 50% in one month. Describe your process to determine where the change is coming from. ↳ Imagine Lululemon has seen a significant drop in checkout completion rate over the past two months. How would you identify where potential customers are dropping out? ↳ Spotify reports a sudden 30% uptick in skipped songs per user. What might be causing this behavior change, and how would you investigate it? ↳ Duolingo sees a 25% increase in daily active users but a decrease in lesson completion rates. How would you go about uncovering the factors influencing these trends? In case you missed it, I shared a framework on how to tackle these metric investigation questions yesterday: https://lnkd.in/ekFfAS-w 👋🏽 I post daily interview tips — follow me for more.

34

12 Comments

Python Data Analysis Tips swarmplot how when why to use a Seaborn swarmplot full how to video walkthrough at DataSimple.education https://lnkd.in/gGs-hpiY Python beginner data analysis tips on how to use seaborn swarmplot(). Learn why you would use this distribution plot versus other types of seaborn plots. learn to analyze your data with the swarmplot() #seaborn #python #dataanalytics #project #dataanalysis #machinelearning #ML #code #teacher #data #dataanalysis #machinelearning #deeplearning #python #learnpython #learndatascience #DataSimple #artificialintelligence #ai #nueralnetwork #datascience #data

3

#Off-Topic Use Python for removing background of Image --Python libraries like Rembg and Pillow can simplify background removal! This code snippet demonstrates how to achieve this: # Install required libraries !pip install rembg pillow from rembg import remove from PIL import Image # Define input and output paths for your image input_path = "in.png" output_path = "out.png" # Open the image input = Image.open(input_path) # Remove the background using Rembg output = remove(input) # Save the output image with a transparent background output.save(output_path) By replacing "in.png" and "out.png" with your image filenames, you can achieve clean cutouts for your graphics.

6

More than never before data quality by design (from proper specification and modelling) and by default (proper control and monitoring on systems) are key, specially for AI as it is highly dependent of accurate and consistent data. More about it with Dan Galavan #datamodelling #datamodeling #AI #bydesign #bydefault

4

Hey all, Carrying on my deep dive into the Data Science Path here in Maven, I wanted to give you a closer look at some of my favorite topics in our Classification Modeling Course. In addition to learning how to fit and Tune KNN, Logistic Regression, Decision Trees, and ensembled trees like Random Forests and GBMs, we worked hard to put together a course that will separate you from the hordes of folks who do model.fit() and call it a day. If you've ever been curious about learning how to supercharge your data skills with powerful classification tools, why not give it a shot today? You won't regret it, I promise. It's a fun, project-based course that will help you speak the language of machine learning and add value in new areas. Hope to see you in the classroom!

34

Are you curious about Exploratory Data Analysis (EDA)? If so, you should definitely check out this document. EDA involves analyzing data sets to highlight their main characteristics, often using visual methods. It's a powerful way to gain insights, identify patterns, test hypotheses, and prepare data for further analysis. Here's what the document covers: 🔴What is EDA and why is it important? 🔴What are the steps involved in EDA and how can you apply them to your data sets? 🔴What are the types of EDA and how can you choose the best one for your data? This guide is perfect for beginners and intermediate learners looking for a comprehensive overview of EDA. Credits to unknown

233

9 Comments