DEV Community: Joseph Morante

Day 2: Delving Deeper into LLMs & Refreshing Machine Learning Basics

Joseph Morante — Wed, 04 Oct 2023 01:02:05 +0000

The journey continues, and today was all about unraveling the intricacies of Large Language Models (LLMs) and brushing up on some good ol' machine learning practices using Python.

Demystifying the Hype around LLMs:

Today, I wrapped up the 'Introduction to Large Language Models' on Coursera. One realization? The hype is real! LLMs are powerful, transformative, and truly at the cutting edge of AI. What stood out to me was how Google's Vertex AI isn't just another platform—it's a holistic solution. From AI solutions to search functionalities, conversations, over a 100 foundational models, and a seamless AI platform, Vertex AI seems poised to change the game.

Back to the Basics with Python:

Switching to Codecademy, the Machine Learning/AI Engineer Career Path took me on a nostalgic journey through Python. Implementing machine learning using Python felt like reuniting with an old friend. The exercises and challenges were a delightful mix of refreshment and new perspectives. The beauty of Python and its powerful libraries never cease to amaze me!

Final Thoughts for the Day:

While it's tempting to rush through the courses, I'm consciously taking the time to grasp, reflect, and practice. Every new lesson is a step closer to mastering AI, and I'm here for the long haul.

Day 2 was a blend of new insights and revisiting core principles. The adventure rolls on, and I'm eager to see where Day 3 takes me.

Stay with me, folks! There's much more to come.

Onto Day 3!

Day 1: LLMs & Codecademy - A Fresh Start in AI

Joseph Morante — Mon, 02 Oct 2023 13:42:30 +0000

"Sucking at something is the first step towards being sorta good at something." - JAKE the DOG (Adventure Time)

The Confusing Beginning:

Stepping into the vast realm of Artificial Intelligence (AI), I found myself at a crossroads—overwhelmed yet excited, with a ton of resources at my fingertips and no idea where to begin. YouTube promised a "simple AI project," but the jargons and the underlying concepts lost me midway.

The Structured Approach:

I decided to shift gears. Coursera and Codecademy have been old friends, so I turned to them for clarity.

Google Cloud on Coursera: I've opted for the course on LLMs, starting with 'Introduction to Large Language Models'. Aim? To demystify LLMs and grasp the potential use-cases.
Machine Learning/AI Engineer Career Path on Codecademy: This is where I'll bridge the gaps in my Machine Learning basics. And the cherry on top? Codecademy aids in building a portfolio with tangible projects.

Looking Ahead:

While this is a promising start, I'm cautious. The infamous "tutorial hell" looms large, and I'm determined to avoid it. My strategy is simple: learn just enough to build, iterate, and then move on to the next challenge.

Day 1 down. A world of possibilities awaits. I'll keep you updated on my progress, missteps, and learnings. Stay tuned!

On to Day 2!

The Start of My Year-long Journey into AI: From Novice to Engineer

Joseph Morante — Fri, 29 Sep 2023 22:53:03 +0000

The world of AI is evolving, and so am I.

LangChain, LLMs, Knowledge Graphs, LlamaIndex (I genuinely thought this was a meme when I first heard the name) — these terms are now omnipresent.

It's official: one can't work in any field without encountering AI and LLMs, as evidenced by Google, OpenAI, and Microsoft to name a few. My ultimate goal is not just to bear the title of an AI Engineer but to truly possess its corresponding skills and responsibilities. However, given the current employer-driven market and recent tech layoffs, I realize the importance of expanding my expertise and executing relevant projects.

I intend to embark on a project-based, self-learning journey, primarily using free resources, and will share updates on my progress. Expect two types of posts: daily snippets providing a quick overview of my activities, and more in-depth monthly reports showcasing significant projects, breakthroughs, or even the setbacks and errors I encounter.

From October 2023 to October 2024, I'm dedicating a segment of my day to this venture. At the journey's conclusion, I plan to compile and share these resources for anyone who might find themselves in a similar situation. Wish me luck!

P.S. I'm not expecting to jump from Data Analyst to AI Engineer within a year but to progressively go from Data Analyst to Data Scientist to AI Engineer. What I expect in this year long journey is to have the foundations necessary to get to the next step of becoming an AI Engineer.

Do you have any insights, suggestions, or similar experiences to share? I'd love to hear from you!

SQL Sunday #1 - Essentials

Joseph Morante — Sun, 27 Aug 2023 18:57:15 +0000

For everyone taking a step to work in the realm of Data Science/Analytics, the first step is to learn SQL (Structured Query Language). As a Data Analyst for about 3 years now while studying for a master in Business Analytics, I've decided as a good review and to improve my knowledge in SQL that I will be posting weekly content on Sundays related to SQL. To start of this first post I'm going to list down 4 essential things to learn when working with SQL.

1. CASE statements
Second to the WHERE clause, I use CASE statements on a daily basis when I need to identify certain values to specific categories. I do advise to be mindful of the order of executions of each condition you create because it may give misleading results. CASE can be used to derive new columns or values based on specific criteria, making it easier to produce complex reports directly from your SQL queries without further processing.

CASE
    WHEN condition1 THEN result1
    WHEN condition2 THEN result2
    WHEN conditionN THEN resultN
    ELSE result
END;

2. WHERE clause
In my opinion, the WHERE clause is like a Swiss army knife. Once you understand how to use it properly, it can solve a wide variety of things whether to pull data based on a certain value in a column to limiting the amount of data I want to see reducing my query cost. Retrieving unnecessary rows from the database consumes resources and is inefficient. The WHERE clause helps you get only what you need, conserving both database processing time and network bandwidth.

SELECT column1, column2, ...
FROM table_name
WHERE condition;

*3. JOINs *
At this point of my career, I can confidently say that more than half of my work would not be possible without the ability to join data tables together. JOINs help with expanding my analysis into answering more complex questions that may need to look into multiple tables at once. Instead of retrieving data from multiple tables separately and then combining them in application code, SQL JOIN allows us to fetch the combined data directly, which is usually more efficient.

SELECT table1.column1,table1.column2,table2.column1,....
FROM table1 
INNER JOIN table2
ON table1.matching_column = table2.matching_column;

4. Aggregate functions (AVG(),SUM(),MAX(),MIN(),COUNT())
Aggregate functions in SQL are indispensable for data analysis and reporting in relational databases. They allow for summarizing, analyzing, and computing data on large datasets to generate meaningful insights. By providing a means to analyze and summarize data, aggregate functions directly support business intelligence and decision-making processes. Business leaders can make informed decisions based on summarized data insights.

Data Analyst: Reality vs Expectations

Joseph Morante — Tue, 25 Oct 2022 11:17:26 +0000

The first rule of data science is: don’t ask how to define data science - Josh Bloom (Azam 2014)

If you are anything like me (I hope not), then you must like to spend your free time looking at job postings on LinkedIn to see what companies in your industry are looking for in a Data Analyst role. Whether for a new job or trying to keep your skills up to date, you'll notice that not all Data Analyst roles are the same. Having worked at 3 different companies ranging from mid-size ad agency, adtech startup and a big fintech I'm going to break 3 expectations of a Data Analyst and explain the reality to each one of them.

You need to know how to code
Starting off with this I'll say right now that the short answer is drum rolls it depends. In my experience, it is not always necessary and depends on what exactly you'll be working on. In my first role out of college as an analyst, I did not write nor used any line of code to do my job. I primarily used Excel and Tableau to work on my assigned tasks. It wasn't until my second and third job, that I got to start using SQL to pull data to build out some projects. My advice would be to defiantly learn SQL, it is a must if you want to be a Data Analyst or even take a step into the world of Data Science. Moving forward, I would recommend learning to code depending where you see yourself going in the long term of your career.
Providing insights and predictions to the company
Now for this, you will be providing insights for your company but not all roles will have you performing predictions. Especially in an entry level role, you will be to focus more on how clients and teams should interpret current data based on the business questions they want answer. It's not until you reach a senior analyst role or transition into a data scientist that developing predictive models becomes part your job or not. Which leads to my final point.
Data Science is the same everywhere
Initially, I thought this to be true when I was in college studying for my degree. I would read and listen to all these books related to Data Science then expect that once I got the role I would be doing those exact things. It wasn't until I started moving around from different companies to realize that each company has their definition to what is considered Data Science, Data Analyst, Data Engineer, and the list goes on. Which is why all jobs in this realm will never be the same. Different companies mean different jobs.

What were some of your expectations that you thought were true when becoming a Data Analyst or working in Data Science in general?