(Dec 2024 - Present)
Software Engineer
Cloudesh Technology
About Me
I am a Software Engineer with a BSc in Computer Science and Engineering and an MSc in Data Analytics from the University of Portsmouth, UK.
With over two years of professional experience across full-stack web development and backend engineering, I have worked at Cloudesh Technology and Impress IT Solutions — delivering production-ready applications, RESTful APIs, and data-driven solutions.
My core stack includes Java, Spring Boot, Laravel, React, and Next.js, backed by solid experience with PostgreSQL, MySQL, and cloud deployment. I am equally comfortable working across the backend, frontend, and data layers.
I am currently open to software engineering roles where I can tackle meaningful challenges, collaborate with driven teams, and keep building things that matter.
LocationLondon, UK
EducationMSc Data Analytics
Emailmdsamialsohan@gmail.com
StatusOpen to Work
Experiences
(April 2021 - August 2022)
Software Engineer
Impress IT Solutions
Skills
PROJECTS
Linkly — Real-Time URL Shortener with Click Analytics
Backend Developer- A production-grade URL shortener demonstrating read-heavy/async-write architecture patterns used in payments, ad-tech, and observability systems, achieving ~12,300 redirects/sec with p99 latency of 9.3ms under k6 load testing and a 100% success rate over 1.11M requests.
- Separated concerns into two paths: a redirect path that resolves via a single Redis cache lookup and returns a 302 in milliseconds bypassing the database entirely, and an analytics path where click events are published fire-and-forget to Redpanda and processed asynchronously by a separate consumer so analytics load never affects redirect performance.
- Generated collision-free, shortest-possible short codes using sequential Base62-encoded counter IDs, and used short-lived negative-cache sentinels to protect against scanner traffic.
- Chose Redpanda over Kafka for its smaller memory footprint while keeping Kafka API compatibility, and used structural `/r/` prefix routing to eliminate ambiguity between short codes and frontend routes.
- Built the frontend in React 19 with Vite, served through Caddy for TLS termination, with a GitHub Actions CI/CD pipeline building ARM64 container images pushed to GHCR and auto-deployed via Watchtower without server-side compilation.
Java 21Spring Boot 3.5PostgreSQLRedisRedpanda / KafkaReact 19ViteCaddyDockerGitHub ActionsGHCRWatchtowerAWSk6
EzDokani — Multi-Tenant SaaS Point-of-Sale System
Full Stack Developer- Architected and developed a multi-tenant SaaS POS platform using Laravel 10 (PHP 8.2) and Next.js 14, enabling tenant-level data isolation through per-tenant MySQL databases dynamically resolved by subdomain using stancl/tenancy v3.
- Designed and implemented 60+ RESTful API endpoints across core business domains, including sales, purchasing, inventory, banking, payroll, customer/vendor ledgers, and 10+ financial reports (P&L, cash flow, stock valuation, customer aging, profitability by product/brand/category).
- Implemented role-based access control (RBAC) with Laravel Sanctum SPA authentication, including an admin approval workflow for salesman-generated sell memos before inventory mutation.
- Built a custom Bangla phonetic input engine from scratch featuring stateful keystroke conversion, auto-conjunct generation, vowel signs.
- Deployed and maintained production infrastructure on a Hetzner VPS using Nginx reverse proxy, PHP-FPM, PM2, and Let’s Encrypt wildcard SSL for *.ezdokani.com, with automated backup pipelines to Backblaze B2.
LaravelNextJSReactNodeCSSTailwind CSSREST APIMySQlPM2NginxPHP-FPMBackblaze B2GitGitHub
TaskFlow — Task Management REST API
Backend Developer- A backend REST API for managing tasks, projects, and team assignments with enforced business rules and full CI/CD.
- Designed a layered architecture (Controller → Service → Repository) with DTOs separating the API contract from JPA entities.
- Implemented status lifecycle rules (TODO → IN_PROGRESS → DONE/CANCELLED), flexible multi-filter search, input validation, and a global exception handler returning structured 404/400/422 responses.
- Wrote integration tests against a real PostgreSQL using Testcontainers and MockMvc tests for the HTTP layer.
- Containerized with a multi-stage Dockerfile, automated tests via GitHub Actions, and deployed continuously to Render.
Java 21Spring Boot 4PostgreSQLHibernate / Spring Data JPAFlywayDockerDocker ComposeJUnit 5TestcontainersMavenGitGitHub ActionsRender
Predicting stock market
Machine Learning Engineer- In this project, I implemented various machine learning algorithms including LSTM, XGBoost, SVR, GRU, and Random Forest (RF) to forecast stock prices for 1 day, 1 week, and 1 month ahead using data from the Dhaka Stock Exchange (DSE).
- The analysis focused on three dominant companies, utilising data collected from January 2021 to January 2024.
- The results from all models were compared to identify the most effective model for predicting stock prices in the DSE.
pythonpandastensorflownumpyLSTMXGBoostSVRGRURF
Educations
2022 - 2024
Master's of Science
Data Analytics
University of Portsmouth, United Kingdom
2016 - 2020
Bachelor of Science
Computer Science and Engineering
Eastern University, Bangladesh
2014 - 2016
Higher Secondary Certificate
Brindaban Govt. College, Habiganj - 3300
2014
Secondary School Certificate
Govt. High School Habiganj