Data Science · Software Engineering

Hi, I'm Sharukh Rahman

Data science student. Software engineer

MSc Data Science at Óbuda University. Previously a .NET software engineer at Encoders, building ERP, trade, and hospital systems. Open to ML, data science, and software roles. Exploring embodied AI, continual learning, and multimodal AI. Projects, notes, and experiments live here.

Get In Touch View Projects

Budapest, Hungary

MSc Data Science, Óbuda

Open to ML / SWE roles

Skills & Technologies

Stack drawn from production .NET work, applied ML projects, and current Data Science coursework.

Programming

PythonTypeScript / JavaScriptC#SQL

ML / Data

PyTorchscikit-learnXGBoost / LightGBMpandas / NumPyOptunaRAG / NLP

Web / Backend

ASP.NET CoreFastAPINestJSReact / Next.jsPostgreSQLMS SQL Server

MLOps / DevOps

MLflowDockerKubernetesAirflowGitHub ActionsTerraform

Experience

Enterprise product modules across ERP, trade, and hospital systems.

Software Engineer (.NET)

Encoders Info Tech Ltd. · Dhaka, Bangladesh

Promoted from Intern (Jul 2023 – Sep 2023)

Oct 2023 – Sep 2025

Owned Finance, Inventory & Sales modules on VitalityCash ERP (bank reconciliation, invoice payments, GL master, stock-out, VAT) with C#, ASP.NET Core MVC, and SQL Server.
Shipped Clarichem loan lifecycle (requisition, stock-validated loan in/out/return, ledgers) plus Excel/PDF exports with EF Core, EPPlus, and iTextSharp.
Built HMS patient ADT workflows, clinical CRUDs, and pharmacy/medicine modules integrated into prescription flows.
Applied onion architecture, unit-of-work, repository–service, and specification patterns with unit tests and stored-procedure optimization.

MSc in Data Science

Óbuda University · Budapest, Hungary

Sep 2025 – Jun 2027 (expected)

CWA 4.56 / 5.00 · Stipendium Hungaricum Scholarship

BSc in Computer Science and Engineering

Green University of Bangladesh · Dhaka, Bangladesh

Jan 2019 – May 2023

CGPA 3.93 / 4.00 · Vice-Chancellor’s & Dean’s Honor

Featured Projects

Applied ML and full-stack systems with public source on GitHub.

RKGM — LLM-Powered Knowledge-Gap Mitigation

Two-phase NLP pipeline that detects prerequisite gaps in research papers via self-consistency prompting, then builds a chronologically ordered learning document from the citation graph — hybrid retrieval (dense + BM25 + RRF + reranking) with a RAGAS faithfulness gate.

RAGLLMNLPsentence-transformersRAGAS

Code

Citi Bike — Demand Forecasting & Station Prioritization

End-to-end ML pipeline for Jersey City Citi Bike: short-term station demand forecasting, imbalance-risk classification, and interpretable station-priority ranking with leakage-safe time-series features and weather signals.

Time seriesscikit-learnpandasFeature engineering

Code

More Projects

Code.Runner — Online Code-Execution Platform

Full-stack NestJS + React platform that executes C, C++, Python, JavaScript, and TypeScript through an asynchronous job pipeline with persistent status tracking and isolated child-process execution.

NestJSReactAsync jobsNode.js

Code

Publications

An Article Recommendation Technique from a Multi-Layer Reference Article Graph for Facilitating Chronological Learning

IEEE STI 2022

A Novel Hybrid Deep Neural Network for Early Detection and Classification of Chicken Diseases

IEEE STI 2023

Certifications

Machine Learning Specialization

Coursera (DeepLearning.AI / Stanford)