Hello everyone! This is my first ever blog on here so if anything is imperfect or my grammar is out of whack, please forgive me. I figured the best way to start off my blogs would be to provide context of what I've worked on in '25 and what I will spend time on the most this year, so here we go...

Had you said to me two years ago that I would be creating my own AI-driven projects in 2026, I wouldn't have believed you. At that time, "machine learning" seemed like a field exclusively for PhDs and engineers from large tech companies. Today, I’m trying out different models, launching small agents, and even training domain-specific models using my own data—all from a 7-year-old Macbook and a dash of curiosity.

In this blog, I’m detailing my approach to AI and ML for this year: my main areas of focus, the tools I’m utilizing, and the methods I’m using to visually monitor my progress to maintain momentum.

So what were you up to last year?

Since I haven't seen or spoken to most you reading in a bit here's a small snippet of what I was up to.

2025 was a very eventful year for me, I exchanged Happy Valley for Philadelphia following my graduation from Penn State and then I started my career at AlphaMeld Corp. where I specialize professionally in the convergence of data and intelligence, concentrating on developing agentic AI models and solid data pipelines.

I spent most of the time last year focused on gaining expertise in MCPs (Model Context Protocol) and Multimodal systems, while on the Data Engineering side I tried to support my knowledge by working with Spark and Snowflake. Away from the office, you would have probably seen me on the mats, having devoted the past year to engaging with Philly’s martial arts scene.

What I’m focusing on in 2026

I’ve focused my learning on three key areas rather than pursuing every new model that is released.

• Agents: Creating compact agents capable of reading documents, invoking tools, and assisting with inquiries.
• Small models: Experimenting with optimized models designed for particular tasks rather than solely relying on large LLMs.
• Deployment: Understanding the fundamentals of model serving and effectively monitoring them

My current stack

Here’s a quick snapshot of the tools I’m using right now to explore AI and ML.

Visualizing my progress

To keep myself consistent, I track my weekly study hours and convert them into an easy-to-read chart that I can quickly glance at, so maybe I can finally be accountable. Here's how the chart looks while im writing this post, from the week of December 21st to now.

I generated this chart with Python and Matplotlib from my own log, exported it. Verdict: Not bad for the first 6 weeks, but need to keep up the pace through out the year.

What’s next?

Next on my roadmap: Use this blog and portfolio website to feature projects with 1. Agentic AI, Multimodal AI and Model Context Protocol (MCP).

I'm already in the dungeon working on these ideas so please be on the look out, I want this blog and website to feature all of my thoughts, experiments and expirences.

I hope you'll stick around and join me on this journey—I'm looking forward to reminisce about these 'early days' in the years to come. Thank you <333

Credits:

Images sourced from: https://www.linkedin.com/school/penn-state-university/ and unsplash.com