Sketchbook — cover page
Python developer, sometimes actor, full-time overthinker.
Powered by coffee, fuelled by curiosity, still figuring the rest out on paper first.
That's the word that fits best — curious, a little introverted, and never quite able to take life too seriously. Born in 2002, I grew up asking questions faster than anyone around me could answer them. These days Artificial Intelligence keeps that curiosity fed.
After my SSC I completed a Diploma in Computer Engineering, while spending just as much energy on debating, acting, poetry and stage performances on the side. Somewhere along the way — between editing silly videos for friends and losing hours to story-driven games — coding stopped being a subject and became the thing I actually wanted to do.
I've paused my academic path for now to explore new experiences, pick up fresh skills, and get ready for the next chapter of my education abroad. My approach is simple: when I love something, I don't have to force it — I just do it, and I learn fast because of it. When the interest isn't there, I'm honest about that too.
It started with jokes — stretching faces in Photoshop and messing around in Premiere Pro to make friends laugh. That playful habit turned into a real interest in creating things.
Then came Python, which is still my strongest brush — I use it daily to break down problems and turn messy ideas into working solutions, with Pandas, PyTorch and TensorFlow opening the door to machine learning and NLP. Git & GitHub became my notebook: every messy draft and polished version, stored and shared.
A mix of machine-learning experiments and small full-stack builds — some polished, some very much still lab notes.
Emotion-detection web app · Python, DistilBERT, Streamlit
Reads what you write, runs it through a DistilBERT model trained on the GoEmotions dataset (27 emotions plus neutral), and tells you the most likely emotion behind it — then suggests games, movies, activities and books to match your mood. The backend, from data prep to training and saving the model, was built from scratch; the frontend leaned on AI assistance to move fast.
Prediction app · Python, Flask
A Flask web app that predicts diabetes risk using the Pima Indians Diabetes Dataset. Users enter health metrics — glucose, blood pressure, BMI, age — and get a prediction back from the trained model.
Smaller GitHub projects · classifiers, regression, web apps
A collection of smaller machine-learning experiments and web builds, including a boosting-based cancer classifier, a cancer regression model, a simple bank management system, and a couple of KNN walkthroughs.
Project summary — The Mind Detective
After finishing a machine learning course, I wanted to go deeper than the coursework — that's where the idea for an emotion detector came from. I chose the GoEmotions dataset as the foundation, a rich set of 27 emotions plus neutral, and trained a DistilBERT model to classify emotion from text, saving the model and tokenizer with Hugging Face's save_pretrained.
The backend — data prep, training, saving the model — was built entirely from scratch. For the frontend, I leaned heavily on ChatGPT to speed things up and polish the interface. Once deployed on Streamlit, the first predictions came back as raw class numbers like class_9; mapping those back to real emotion names let the app finally speak plain language: "You are feeling joy."
The result is a clean, simple web app that reads text, understands it through a trained model, and responds with the most likely emotion behind it — a mix of real backend work and some handy AI-powered assistance on the front end.
One of my more unusual habits — I walk around explaining things out loud. If I see something online I disagree with, I instantly build the counter-argument in my head. It keeps my thinking sharp, playful and a little bit combative.
Games filled with depth, creativity and symbolism are more than entertainment to me — they're interactive stories that inspire imagination and emotion. I find myself picking apart their hidden layers long after I've put the controller down.
My journey with debate began at 12, through Fulkuri, a national-level organization. Competing on that stage gave me the confidence to express ideas clearly, sharpened my logical thinking, and gave me a genuine taste for national-level competition.
I started acting at 10, inspired by Shah Rukh Khan and later Barry Evans, and performed regularly from school stages up to national platforms. The most memorable moment came on August 12, 2022 — winning Best Actor at the National Theatre Competition organized by SOSAS, and receiving the award from actor and activist Ilias Kanchan.
Honoured for a stage performance recognised at the national level.
Recognised for outstanding dramatic expression and stage presence.
Promoted to Opposition Leader, strengthening debate and organisational leadership skills.
Recognised as exceptionally skilled in field work and parades — discipline, teamwork and execution.
Took on debate and leadership responsibilities as Opposition Whip.
Served as Shongshodiyo Shodossho, practising leadership and debate at a very young age.
Several school-year awards for consistent performance and dedication.
"These are just material things. The real award is the experience I gained along the way."
Kratos said this once, and it captures exactly how I feel about the word "sorry." If you've messed something up in my life, don't drown me in apologies — fix it, or leave me the space to fix it myself. Repeating "sorry" only keeps the mistake alive in my head. What I actually need is peace, and the room to heal in my own way — forgiveness only becomes real once I feel it, not once it's been said enough times.
If I say I want to reach Amsterdam, what happens the moment I get there? Instead, pick a direction — let your goal be moving west simply because you love going west. Every step toward it becomes a success on its own, and even after "arriving" there's always more west to explore. Love the journey, not just the destination, and you'll find better experiences — and better company — along the way.
Fiction has always been about the experience. When film first showed audiences a train pulling into a station, people reportedly panicked — a level of immersion books alone couldn't touch. For a long time film held that crown. Now I think games have taken it: they don't just tell you a story, they let you live it, choose it, and shape how it ends.
Discipline needs seriousness — the maturity to keep emotion from derailing a goal. But when your emotions actually support the goal you're chasing, seriousness quietly disappears. You stop forcing it. You think about it all day, do it without hesitation, and genuinely enjoy every part of it. That's when passion turns effortless.
Dogs are amazing — genuinely my third favourite animal — but horses take the top spot for me. They've carried people into war, stood on the battlefield, and helped bring the injured home. The trust built with a horse in a life-and-death moment sits on a different level entirely. Mostly, though, I just love horses. No further logic required.
Aspiring Python developer with a strong foundation in programming and a growing interest in machine learning — looking to contribute to real-world projects and grow into a proficient AI & data-driven developer.
Brain Station 23 — Python & Django backend development, since July 2023.
Diploma in Engineering, Munshiganj Polytechnic Institute (2019–2023) · SSC, Hazi Pandey Ali High School (2017–2019).
Bengali (native) · English (advanced)