Mohd Kamran Tahir

A comprehensive SaaS application to enhance the review system of eCommerce sites, and, as for the tech stack, Go with Chi and Huma on the back end, Python for the LLM-related pipelines, and React on the front end are used. The app includes features like autocomplete to make writing reviews easier, spam checking to ensure content integrity, and large language models (LLMs) support to improve user interaction.

The application integrates quality assurance (QA) to maintain high standards and offers an easy user funnel for a seamless review submission process. Additionally, it provides tools for analyzing user thoughts, automatically extracts key topics from reviews, and reports on any issues, helping to gain valuable insights and address customer concerns efficiently.

A full-stack app created for avid readers to scrape, save, and translate Chinese novels. This was my 1st big project, mainly done in Node.js, TypeScript, and React.

Users can review per novel, comment per chapter, upvote, downvote, etc.

Turn unstructured documents into structured intelligence with enterprise-grade accuracy. LeapOCR blends computer vision with language models. The platform uses a schema-first OCR approach for messy, real-world documents, transforming PDFs, scans, photos, and multilingual paperwork into Markdown for review or JSON that already fits downstream schemas.

Tafsir API is a public REST-style data API that provides free access to Quran tafsir content for websites and applications. I built it to be simple to consume, fast to serve through static/CDN-friendly endpoints, and easy for developers to self-host. The project supports multilingual tafsir editions, no-rate-limit access patterns, clear endpoint versioning, and contribution paths for expanding the data set.

uids-postgres is a PostgreSQL extension for generating modern unique identifiers directly inside the database. I built it in Rust to support UUID v6/v7, NanoID, KSUID, ULID, TypeID, Timeflake, PushID, and CUID2 workflows, giving teams database-native ID generation for distributed systems. The project focuses on ergonomics, data integrity, predictable SQL functions, and scalable identifier strategies for high-volume applications.

nextjs_approuter_electron is a public TypeScript template for building desktop apps with Electron and the Next.js App Router. I created it to combine Electron with SSR and React Server Components, giving developers a practical starting point for modern desktop apps without losing the productivity of the Next.js ecosystem. The project demonstrates cross-platform app architecture, packaging concerns, and full-stack UI patterns.