I am a Compiler Engineer & Systems Architect specializing in native toolchains, C++23 named modules, declarative UI domain-specific languages (DSLs), and binary metadata generation. I architect and build high-performance systems from the ground up: from native compilers that transform Windows SDK headers and declarative UI markup directly into ECMA-335 .winmd binaries in seconds, to bounded multi-step agentic AI systems, enterprise multi-tenant ERP platforms in Laravel & Next.js, and RAG pipelines.

A unified next-generation Windows toolchain engineered in pure C++23 named modules (.ixx), eliminating legacy MSBuild overhead, fragile scraping pipelines, and thousands of lines of manual configuration.

Systems & Compilers

Languages

AI & Machine Learning

Frameworks & Web Ecosystem

Databases, DevOps & Cloud

Compiler Architect & Systems Engineer | Open-Source Creator  (2020 – Present)

• WinverseC-Compiler — Win32 & WDK Driver Metadata Compiler (C++23 Modules)

  • Designed and built a high-performance Win32 & WDK Driver API compiler from scratch in C++23 modules with MSVC.
  • Implemented an in-memory ECMA-335 binary writer synthesizing valid PE32+ assemblies with #~, #Strings, #Blob, and #GUID metadata streams, completely eliminating .NET runtime and Roslyn dependencies.
  • Engineered a recursive-descent parser parsing Windows SDK (26100+) headers (um/, km/, shared/) with architecture tagging (X86, X64, Arm64) in ~30 seconds (replacing Microsoft's 190,000-line ClangSharp/Roslyn pipeline).
  • Extracted full SAL annotations into BufferSemanticsAttribute, decoded IOCTL CTL_CODE macros into structured attributes, and classified 6 handle taxonomy types across 64 custom attributes.

• WinUIC-Compiler — Declarative UI DSL & ECMA-335 Compiler for WinUI 3

  • Architected a declarative UI domain-specific language (.winui) for WinUI 3 / Windows App SDK compiling into standardized ECMA-335 .winmd metadata and packaged PRI resources (resources.pri).
  • Achieved the flagship architectural pillar of Zero XAML, Zero IDL, Zero Code-Behind, and Zero manual UI code across any language projection.
  • Built an in-RAM resource indexer calling Windows SDK mrmsupport.dll native APIs directly, accompanied by a zero-dependency binary PRI fallback encoder (mrm_pri2).
  • Implemented a Pratt parser expression engine with deterministic operator precedence, reactive state observers (@state, @binding), immersive dark titlebar attributes, and dynamic third-party control registries.

• Xaml2WinUIC-Compiler — WinUI 3 XAML to .winui Transpiler

  • Built a native C++23 transpiler converting WinUI 3 / Windows App SDK XAML projects and code-behind into 100% equivalent .winui DSL files.
  • Resolved dual-language code-behind dialects: C# partial classes, x:Bind targets, and C++/WinRT / C++/CX ref class / ^ handle declarations.
  • Preserved verbatim layouts and resource fragments via .raw() modifiers with zero design loss; verified against the complete WinUI 3 Gallery corpus with 190/190 XAML files converted with 0 failures and 0 warnings.

• CppWinC-Compiler — Unified WinRT, WinUI 3, Win32 & Driver (WDK) C++23 Projection Compiler

  • Implemented a unified projection and component generator emitting pure C++23 named modules (.ixx) from ECMA-335 .winmd metadata for WinUI 3 & Windows App SDK (with dedicated -winui auto-discovery of package metadata, activation factories, and runtime DLLs), standard WinRT (Windows.*), Win32 (Windows.Win32.*), and Driver APIs (Windows.Wdk.*).
  • Completely eliminated header files, precompiled headers (pch.h), and IFC build order conflicts, enabling direct import std;.
  • Engineered multithreaded codegen across all CPU cores, automatic monolith namespace partitioning (splitting 94K+ methods in Microsoft.UI.Xaml into balanced sub-modules), and base-chain inheritance projection for up to 5,000 methods per type.

• Laravel-PostgreSQL-Multi-School-Management-System — Enterprise Multi-Tenant Educational ERP (Laravel)

  • Architected a multi-tenant educational management system in Laravel supporting isolated tenant database environments via global query scopes and custom tenant middleware.
  • Implemented granular 4-role RBAC with Laravel Gates and Policies across Admin, Teacher, Parent, and Student roles with automated school-specific GPA calculation engine and optimized eager loading to eliminate N+1 queries.
  • Integrated high-performance React/Vite frontends consuming RESTful Laravel API resources.

• Django-PostgreSQL-Nextjs-AI-Research-Agent (CodeFusion): Engineered an autonomous AI codebase exploration agent powered by Google Gemini 2.0 Flash function calling with a 24-round bounded tool loop and crash-safe PostgreSQL audit trail.

• Nextjs-VentureForgeAI-Client (VentureForge AI): Full-stack AI startup incubator and financial validation portal with Groq Llama 3.3 70B evaluation engine, CFO AI financial audit dashboard, and Llama 3.2 11B Vision pitch deck analysis.

• Nextjs-PostgreSQL-MongoDB-Collaborative-Team-Hub (Team Hub): Architected a real-time collaborative workspace with dual-database architecture (PostgreSQL + MongoDB), Socket.IO channels, and transparent Axios JWT refresh interceptors.

• Nextjs-AI-PDF-SaaS (Bookified): Shipped a production RAG voice & chat SaaS using OpenAI Whisper STT, TTS, LangChain, and Vercel AI SDK streaming.

WebFlow Developer & Technical Integrator : Flow Seek  (2020 – 2023)

• Designed and delivered 20+ responsive web applications and landing pages, translating complex Figma wireframes into production builds with 90+ Core Web Vitals scores.
• Architected reusable Webflow CMS architectures for non-technical stakeholders, streamlining post-launch content management.
• Extended native capabilities with custom JavaScript embeds and third-party integrations (Stripe, Google Analytics, Mailchimp, Zapier).

• LL.B (Hon's) in Al-Fiqh and Law — Islamic University, Kushtia-Jhenaidah (2019 – 2023) — CGPA: 3.28 / 4.00
• Higher Secondary Certificate (H.S.C), Science — Nachole Govt. College, Rajshahi Board (2018) — GPA: 4.58 / 5.00
• Secondary School Certificate (S.S.C), Science — Maktapur High School, Rajshahi Board (2015) — GPA: 5.00 / 5.00

English (Fluent) · Bengali (Native) · Hindi (Working Knowledge)