Software Engineer — Dhaka

Sajidur Rahman Sajid

I build high-throughput backend systems and production ML pipelines — engineered to be fast, reliable, and worth trusting.

View projects Download CV

CGPA / 4.00

tests written

thesis dataset

Dean's Award

zsh — portfolio

~/src

Python

FastAPI

Node.js

Express

NestJS

PostgreSQL

SQLAlchemy

Redis Streams

BullMQ

Docker

REST

PyTorch

CatBoost

Scikit-learn

Hugging Face

Pandas

NumPy

Transformers

OpenCV

SHAP / LIME

React

Next.js

TypeScript

Tailwind CSS

Vite

Redux

Framer Motion

MongoDB

SQLite

Supabase

Cloudflare R2

Qdrant

Neo4j

Solidity

Foundry

OpenZeppelin

Viem

Turnkey KMS

Git

pytest

CI / CD

Vercel

Linux

Npm Workspaces

Python

FastAPI

Node.js

Express

NestJS

PostgreSQL

SQLAlchemy

Redis Streams

BullMQ

Docker

REST

PyTorch

CatBoost

Scikit-learn

Hugging Face

Pandas

NumPy

Transformers

OpenCV

SHAP / LIME

React

Next.js

TypeScript

Tailwind CSS

Vite

Redux

Framer Motion

MongoDB

SQLite

Supabase

Cloudflare R2

Qdrant

Neo4j

Solidity

Foundry

OpenZeppelin

Viem

Turnkey KMS

Git

pytest

CI / CD

Vercel

Linux

Npm Workspaces

[ 01 ] — cortex

Peek inside the machine.

This isn't an about page. It's the console to the engineer's brain — run a command, poke around, and learn the person by exploring the system.

cortex — self-inspection consolev2.4.1

online0xfivb

type a command · ↑ history

awaiting input_

run helpor click a command above to explore what's inside.

history—

cortex loading_

[ 02 ] — Projects

Systems that earn trust.

A curated collection of software engineering, AI, computer vision, graphics, database, and production systems I've built.

View all projects

InvoicePilot — AI-powered invoice processing platform that extracts, validates, and analyzes invoices while detecting duplicates and anomalies to streamline financial document workflows.

ProfessionalProduction

Featured

InvoicePilot

AI-powered invoice processing platform that extracts, validates, and analyzes invoices while detecting duplicates and anomalies to streamline financial document workflows.

FastAPIFastifyBullMQPostgreSQLSolidity

Case Study

LedgerTurf — Real-time turf booking platform that lets players discover and reserve grounds on a map while owners publish and manage slots with overlapping-time protection.

Professional

Featured

LedgerTurf

Real-time turf booking platform that lets players discover and reserve grounds on a map while owners publish and manage slots with overlapping-time protection.

Next.jsTypeScriptMapbox

GitHub Live Demo Case Study

CaseVault GraphRAG — Privacy-first legal research workspace that ingests case documents, ranks search results by relevance, and provides AI-generated summaries with verifiable citations.

ProfessionalProduction

Featured

CaseVault GraphRAG

Privacy-first legal research workspace that ingests case documents, ranks search results by relevance, and provides AI-generated summaries with verifiable citations.

GraphRAGNeo4jQdrantFastAPILLMs

Case Study

Natural Language Processing

FinBERT Financial Sentiment Analysis

Comparative study proving domain-tuned transformers outperform generic models on financial sentiment, with a full fine-tuning-to-evaluation pipeline in a single notebook.

BERTTransformersPyTorchNLP

GitHub

Desktop

CodingVibes Learning Platform

Interactive Java learning platform featuring courses, quizzes, and progress tracking built around a clean event-driven architecture with persistent state.

JavaSwingMySQL

GitHub

Computer Vision and Pattern Recognition

Face Recognition System

Real-time face identification pipeline that detects faces in video, trains embeddings on a known set, and identifies people live — from dataset to inference in one reproducible notebook.

PythonOpenCVFaceNet

GitHub

Desktop

SparkPowerhouse Gym Desktop

Windows gym management application separating member and admin workflows for memberships, payments, and daily records in a role-aware system.

C#WinFormsSQL Server

GitHub

Graphics

3D Procedural City Generator

Procedurally generated 3D city with dynamic day-night lighting, rain and snow effects, and functioning traffic-light logic in a pure graphics showcase.

C++OpenGLSFML

GitHub

Web

SparkPowerhouse Gym Web

Web-based gym management system with distinct member and admin areas for class bookings, memberships, and billing in one session-driven application.

PHPJavaScriptCSSMySQL

GitHub

Desktop

Employee & Family Registry

Employee registry with family relationship trees, full-text search, and on-demand PDF CV and list exports in one polished API-driven workspace.

C#SQL Server

GitHub

Creative

Interactive Wedding Invitation

Cinematic digital wedding invitation with RSVP, live countdown, polaroid gallery, and venue maps — designed and deployed as a live experience.

Next.jsTailwind CSSFramer Motion

GitHub Live Demo

Database

Online Bookstore Database

Fully normalized relational design for an online bookstore mapping every entity and dependency, ready to run catalog, orders, and inventory queries.

SQLNormalizationER Diagram

GitHub

[ 03 ] — Research

Featured thesis

A hybrid ensemble trained on 2,237 student surveys detects cognitive impairment better than any single model.

Thesis · ML / Deep Learning

NeuroScreen

A CatBoost + ANN hybrid ensemble that detects cognitive impairment in insomniac university students — 95.20% accuracy, 0.982 ROC-AUC.

Natural Language ProcessingMachine Learning

Twitter Sentiment Analysis with TF-IDF and Multinomial Naïve Bayes

Natural Language Processing

Course research on the Sentiment140 corpus. 1.6M tweets are balanced down to 200k (100k positive + 100k negative), run through a deep preprocessing chain — lowercasing, URL/@/# removal, punctuation stripping, stopword removal, WordNet synonym substitution, Porter stemming, and lemmatization — then vectorized with TF-IDF and classified by Multinomial Naïve Bayes on a stratified 80/20 split.

Machine LearningPredictive AnalyticsBusiness Intelligence

Early Warning Model for High-Value Customer Drop-Off

Data Science · Customer Analytics

Course research in Data Science. Cleans 541,909 transactions into a 4,338-customer RFM matrix, segments customers with unsupervised clustering compared across three algorithms, then builds a supervised early-warning classifier on month-over-month RFM decay to flag high-value customers at risk of churn.

early-warning signals

0.788

ROC-AUC

Computer VisionDeep LearningTransfer Learning

Improving Generalization of Image-Based Water Turbidity Classification Using Fine-Tuning with Small Real-World Datasets

Computer Vision & Pattern Recognition

Course paper that tackles image-based turbidity classification with a small custom dataset of 300+ photos taken on three different smartphones in natural, uncontrolled lighting. An ImageNet-pretrained EfficientNet-B0 with partial fine-tuning learns to separate Low / Medium / High turbidity, and the paper candidly reports the limits of doing so from photos alone.

Natural Language ProcessingBERTTransformersExplainable AI

Explainable Bangla Toxic Comment Detection using BanglaBERT with SHAP and LIME

Data Warehouse & Data Mining

Data-mining research (IEEE-format paper) that fine-tunes BanglaBERT on Bengali toxicity data using 5-fold stratified cross-validation and adds an explainability layer — SHAP and LIME reveal which words drove each toxic / non-toxic decision, tackling the black-box problem in low-resource Bangla NLP.

[ 04 ] — Experience