I am an AI/ML Engineer, Full-Stack Developer, and Data Scientist. Currently pursuing my Master of Science in Data Science at Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU) in Germany.

My expertise is rooted in building scalable AI-driven applications, real-time analytics engines, and state-of-the-art modern web platforms. I am deeply passionate about pushing the boundaries of machine learning and solving complex problems with beautifully crafted full-stack applications.

• TheBrainyInsightsReplatform: A complete replatforming project for The Brainy Insights, architected using TypeScript for high scalability.
• IAEC: Web platform repository associated with IAEC Consultants.
• Admin_Dashboard: A robust administrative dashboard interface.

• AI_Pipeline: End-to-end artificial intelligence pipeline architectures.
• Drone_Challenge: Machine learning and computer vision solution developed for the Flipkart-grid challenge using Jupyter Notebooks.
• ResumeParser: Natural Language Processing (NLP) tool developed to extract and structure key information from resumes efficiently.
• AI-Wardrobe-Management: Computer vision and recommendation system integration designed for wardrobe management.
• Machine_Learning & DL_Assignments: Comprehensive collections of deep learning models, algorithms, and academic assignments from FAU.

• DinoCLI: A robust Python-based Command Line Interface (CLI) application.

I am an active contributor to the open-source community, regularly committing to top-tier machine learning, systems, and full-stack repositories. To date, I've contributed over 70 PRs to major open-source frameworks.

• composable-data-stack: Authored core regression tests preventing raw secret leakage; added docs and CLI examples (PR #24, PR #40).
• Solar-Wanderer: Integrated historically accurate traverse coordinates for Mars rovers (Perseverance, Curiosity, Opportunity, Spirit, Zhurong) into the scene builder (PR #56).
• lens: Implemented ActivityStreams and OriginTrail Knowledge Asset fixtures (PR #12).
• Travel-Itinerary-Parser: Fixed complex date parsing edge cases (PR #8).
• argus: Authored advanced time series forecasting research documentation (PR #63).
• kana-dojo: Expanded learning database with Japanese Haiku fixtures (PR #21438).
• neural-quantum-consciousness-framework: Shipped core bug fixes and features (PR #12).

• vLLM: Spearheaded a massive set of optimizations and fixes across 12 distinct PR campaigns, including async engine memory leak resolution (PR #46343), Qwen3.5 MoE architecture support (PR #46297), and RTX 5090/SM101 FP4 GEMM init fallbacks (PR #46320).
• HuggingFace Transformers: Led 11 PRs solving critical issues such as FSDP gradient checkpointing mismatch (PR #46804), Pipeline text generation OOM (PR #46805), and Tokenizer Fast special token drops (PR #46801).
• LangChain: Shipped fixes for convert_runnable_to_tool PII middleware exports and routing issues across multiple PRs (PR #38752, PR #38356, PR #38757).
• PyTorch Lightning: Resolved FSDP device mesh initialization on CPU (PR #21820) and ModelCheckpoint deletion issues (PR #21819).
• HuggingFace PEFT: Optimized CUDA memory usage for k-bit training and added optimum-quanto support (PR #3347, PR #3348).
• HuggingFace Diffusers: Fixed CLIP text model prefixes for FLUX LoRA (PR #14035).
• LiteLLM: Added xai-oauth support and GPT 5.6 model routing features (PR #33815, PR #33814).
• Model Context Protocol SDK: Contributed core framework improvements (PR #2938, PR #2940).
• LLVM Project: Active pipeline updates and fixes (PR #204954).