Hello, I'm

Nahid Hassan

Product Intern @ Pathao

I work at the intersection of product thinking, student-facing software, and applied AI. My recent work spans ride-product execution at Pathao, academic tooling, plagiarism detection, text clustering, and communication-heavy technical writing.

Projects

07

Academic and applied builds spanning product workflows, web platforms, NLP, graphics, and operating systems.

Students Tutored

500+

Supported across Discrete Mathematics and Numerical Methods through structured academic guidance.

Certifications

05

Selected learning credentials across AI, communication, strategy, and professional development.

Portrait of Nahid Hassan

Current Focus

Product execution for mobility systems, practical NLP projects, and educational tools that solve real friction instead of adding novelty.

PathaoBRACUNLP

Experience

Work that blends product, communication, and execution.

My work has moved between product management, teaching, writing, and execution-heavy collaboration, with the same emphasis on clarity and follow-through.

Jan 2026 - Present

Product Management Intern

Full-time

Pathao Ltd. (Rides Team) | Dhaka, Bangladesh

Contributing to ride-product planning, feedback-driven improvements, and feature execution for Pathao's Rides platform.

  • Worked on features including Intercity Ride Visibility, CNG Pricing Fix, and a corrective system informed by ongoing and post-trip feedback.
  • Collaborated with engineering and design teams in Jira and Figma to define user flows, write user stories, and track implementation progress.
  • Analyzed user behavior with SQL and Firebase to identify issues and support product decisions with real usage data.

Jun 2024 - Sep 2025

Undergraduate Teaching Assistant / Student Tutor

Part-time

BRAC University | Dhaka, Bangladesh

Supported students in Discrete Mathematics and Numerical Methods through tutoring, assessment feedback, and close faculty coordination.

  • Guided 500+ students across four consecutive semesters in CSE230 and CSE330.
  • Helped reduce learning gaps by explaining technical concepts clearly and reinforcing problem-solving habits.
  • Assisted instructors in checking quizzes and assignments while giving constructive feedback to students.

Mar 2024 - Jun 2025

Contributing Writer and Journalist

Part-time

BracU Express | BRAC University

View author archive

Covered campus stories, wrote feature pieces, and developed a stronger voice for audience-focused storytelling.

  • Produced articles and opinion pieces read by more than 1,000 students across BracU platforms.
  • Used interviews with students and faculty to build well-grounded stories with clear human context.
  • Strengthened editorial clarity and consistency through repeated publication and feedback cycles.

Jul 2021 - Oct 2021

Content Writer

Part-time

ATZ Technologies | Dhaka, Bangladesh

Wrote SEO-focused web content and helped strengthen the company's digital presence.

  • Created SEO-friendly blog posts and web content using keyword research and WordPress.
  • Contributed to a 15% increase in website traffic during the internship period.
  • Learned how to write with both search intent and readability in mind.

Projects

Academic and applied builds with real technical depth.

These projects cover the kinds of problems I keep returning to: student workflows, text understanding, product clarity, and systems that need to be usable beyond the prototype stage.

BracU Diary
20251 min read

BracU Diary

A unified student platform built to centralize academic tools, course resources, events, and AI-assisted support for BRAC University students.

Next.jsTypeScriptFastAPIMongoDB

Role Full-Stack Developer

Outcome Combined student resources, event information, and an AI support layer into one structured academic platform.

AI-Powered Plagiarism Checker
20251 min read

AI-Powered Plagiarism Checker

An academic integrity tool that analyzes text similarity using Python NLP models and presents results through a clean Laravel interface.

LaravelSQLitePythonNLP

Role Full-Stack Developer

Outcome Created a faster and more accessible workflow for identifying duplicated content and presenting clear similarity results.

Reflectix
Reflectix alternate preview
20241 min read

Reflectix

A multi-level 2D DX-ball style game built with Python and OpenGL to explore reflections, collisions, and progressive gameplay.

PythonOpenGLComputer Graphics

Role Game Developer

Outcome Translated graphics concepts into an interactive multi-level game with collision logic and increasing difficulty.

Multi-Class Text Classification System
20251 min read

Multi-Class Text Classification System

A large-scale text classification project comparing machine learning and deep learning models across a 340K+ question-answer dataset.

PythonTensorFlowScikit-learnNLP

Role ML Developer

Outcome Benchmarked multiple modeling strategies and achieved the strongest result with a BiLSTM at roughly 72.5% accuracy.

A-DEC++ Text Clustering
20251 min read

A-DEC++ Text Clustering

An adaptive deep embedding clustering framework for unsupervised text clustering using Sentence-BERT representations and custom clustering refinements.

PythonSentence-BERTPyTorchScikit-learn

Role ML Research Project

Outcome Benchmarked a custom A-DEC++ clustering approach on DBpedia and 20 Newsgroups using multiple clustering metrics.

CSE321 UNIX Shell
20241 min read

CSE321 UNIX Shell

A custom UNIX shell written in C for BRAC University's CSE321 Operating Systems course, supporting redirection, piping, sequencing, and signal handling.

CUNIXShellSystems Programming

Role Systems Programming Project

Outcome Implemented a working shell that handled command parsing, piping, redirection, command history, and CTRL+C behavior.

CSE321 VSFS Consistency Checker
20241 min read

CSE321 VSFS Consistency Checker

A file system consistency checker for a custom VSFS image built for BRAC University's CSE321 Operating Systems course.

CFile SystemsVSFSOperating Systems

Role Systems Programming Project

Outcome Built a checker that validated superblock metadata, bitmaps, inode references, duplicate blocks, and bad block conditions in a corrupted file system image.

Research

Undergraduate thesis on runtime-adaptive pruning for large language models.

My thesis work focuses on SPRINT, a sensitivity-guided framework for selecting structural pruning decisions at inference time using oracle labels, a learned router, and reinforcement learning.

Thesis

Undergraduate Thesis: SPRINT - Sensitivity-guided Pruning for Inference-Time Adaptation of LLMs

June 2025 - April 2026 | Completed

LLMsStructural PruningReinforcement LearningAdaptive Inference

A runtime-adaptive framework for selecting structural pruning intensity per prompt using oracle sensitivity labels, a learned router, and a DDQN controller. Instead of compressing the model once and using the same structure for every prompt, the thesis explores how runtime signals can help choose different pruning actions based on complexity, hardware state, and quality-speed tradeoffs.

The core pipeline combines oracle sensitivity labels, a learned complexity router, and a DDQN controller so the system can decide when to skip layers or prune attention heads while keeping degradation bounded.

  • Built on a 10,000-prompt multi-domain benchmark spanning math, code, language modeling, and QA.
  • Designed to deliver structural speed gains that are visible at inference time, not just sparse weights on paper.
Open in new tab

Skills

The product, technical, and communication stack behind the work.

The strongest part of my work sits where product judgment, technical implementation, and communication support each other.

Product and Analysis

Turning user feedback, metrics, and requirements into clearer product decisions.

Product ManagementRequirement GatheringUser StoriesRoadmappingJiraFigmaSQLFirebase

Software and ML

Building academic and product-facing systems across web, data, and machine learning workflows.

Next.jsTypeScriptFastAPIMongoDBLaravelSQLitePythonTensorFlowScikit-learnJavaC

Writing and Communication

Explaining technical ideas clearly through tutoring, journalism, and structured content work.

Technical WritingContent WritingSEO WritingTutoringPublic CommunicationLeadershipCanvaWordPressPower BITrello

Certifications

Selected certifications that sharpen communication, strategy, and applied AI work.

Focused learning credentials that reinforce how I approach AI systems, communication, and strategic thinking.

Yale University

The Science of Well-Being

A foundational course on behavior, habit design, and personal effectiveness.

LinkedIn Learning

Essential Career Soft Skills

Focused on communication, professionalism, and practical workplace judgment.

DataCamp

AI Engineer for Data Scientists Associate

Built around applied machine learning workflows and AI implementation skills.

LinkedIn Learning

Strategic Thinking

Structured approaches to prioritization, decision-making, and long-range planning.

LinkedIn Learning

Elevate Your Public Speaking

Training in presentation clarity, confidence, and audience-oriented delivery.

Writing

Writing across campus journalism, culture, and literary publication.

I treat writing as part of how I think: reporting carefully, structuring ideas cleanly, and making the final piece feel human instead of generic.

BracU ExpressJune 8, 20251 min read

In Conversation with Tamara Abed: A Legacy of Resonance

An interview feature reflecting on Sir Fazle Hasan Abed's principles, social vision, and continuing influence.

InterviewBRACU ExpressSpotlight

Published on BracU Express. This feature captures reflections on leadership, integrity, and BRAC's social legacy through a conversation with Tamara Hasan Abed.

BracU ExpressJune 2, 20251 min read

In Conversation with Shameran Abed: A Remembrance in Legacy

A feature on remembrance, BRAC's mission, and how Sir Fazle Hasan Abed's legacy continues through institutions and values.

InterviewBRACU ExpressLegacy

Published on BracU Express. The piece centers on memory, institution-building, and the broader civic philosophy that shaped BRAC and BRAC University.

BracU ExpressApril 28, 20251 min read

The Dojo Reopens: What to Expect from the New Karate Kid Film

An entertainment feature previewing Karate Kid: Legends through its generational crossover, themes, and audience expectations.

EntertainmentBRACU ExpressFilm

Published on BracU Express. This article looks at the new Karate Kid film through franchise history, character continuity, and what long-time fans might expect.

The Daily StarApril 5, 20241 min read

'Soul': Sehri Tales selections, Day 25

A reflective Sehri Tales selection about burnout, disorientation, and regaining a sense of presence through small moments of stillness.

Star LiteratureSehri TalesFlash FictionThe Daily Star

Published in The Daily Star / Star Literature. This piece is one of the more introspective selections from my Sehri Tales work, centered on exhaustion, memory, and the search for direction.

About

A builder who likes systems that stay useful and legible.

The common thread in my work is straightforward: take a messy problem, structure it well, and make the outcome easier for someone else to use.

Nahid Hassan

What I care about

I learn fastest when I have to build, explain, and iterate at the same time. That is why product work, tutoring, journalism, and software projects all feel connected in my career rather than separate tracks.

I am especially interested in how NLP and LLM-assisted systems can improve educational and user-facing experiences without losing trust, structure, or day-to-day usefulness.

Current direction

Right now I am focused on product roles where user feedback, analytics, experimentation, and execution all matter, while continuing to build AI and NLP systems that are useful outside the lab.

Contact

Open to internships, collaborations, and meaningful conversations.

If you are working on product, education, student platforms, or applied AI and think there is a fit, I would be glad to connect.

Availability

Available for internships, project collaboration, writing, and community work.

Email is the fastest way to reach me. If it helps, send a short note on the context, timeline, and kind of collaboration you have in mind.