Building production-grade AI/ML systems, LLM applications, intelligent developer tools, and scalable backend services.
I specialize in machine learning, deep learning, NLP, computer vision, generative AI, and LLM-powered applications — taking systems from model to deployable product. Strong secondary expertise in Python/FastAPI and ASP.NET Core/.NET backend engineering.
Machine Learning Engineer focused on systems-level ML
I'm a Machine Learning Engineer who builds production-grade ML systems from the ground up. My flagship project, VectorForgeML, is a complete ML framework written from scratch in C++ — featuring BLAS/LAPACK hardware acceleration, OpenMP parallelism, and zero-copy R integration.
I'm an AI/ML engineer focused on building practical, production-oriented intelligent systems. My work spans machine learning, deep learning, NLP, computer vision, generative AI, LLM applications, RAG, and AI-powered developer tools.
I also have strong backend engineering experience with Python/FastAPI and ASP.NET Core/C#, allowing me to take AI systems from models and APIs to fully deployable applications.
My technical differentiator is depth in AI/ML combined with the ability to build the backend infrastructure that ships it — whether in Python, C#, or C++.
Bachelor of Computer Applications (BCA)
Takshashila Mahavidyalaya, Amravati · Affiliated: Sant Gadge Baba Amravati University
CGPA: 8.09 / 10 · Completed
AI-powered Unity NPC that understands voice commands, answers programming questions, and responds with speech, animations, and actions such as run, stop, and wave.
Core technologies and tools I work with
Selected projects from 25+ ML/AI builds
Production-grade AI security tool combining static analysis with LLM-based semantic vulnerability detection
AI-powered repository security analyzer combining deterministic static analysis with Llama 3.3 70B semantic vulnerability detection to identify risks, explain findings, generate remediation suggestions, and create patch-ready fixes. Supports GitHub, GitLab, ZIP uploads, and local projects. Exports SARIF, HTML, JSON, CSV, PDF, and Markdown reports with unified diff patches.
AI-powered repository security platform that combines static analysis, LLM-based semantic vulnerability detection, cross-file reasoning, and automated security reporting for GitHub, GitLab, ZIP archives, and local projects.
Dockerized Personal RAG platform that builds AI assistants from GitHub, portfolios, resumes, blogs, and custom URLs using ChromaDB semantic retrieval.
Deep learning news summarizer achieving ROUGE-1: 71.43% and BERTScore: 0.93 using PRIMERA & LED architecture.
AI-powered resume screening system processing 1000+ resumes using TF-IDF vectorization for candidate-job matching.
End-to-end ML pipeline for predicting employee attrition with feature engineering and API-ready architecture.
AI-powered YouTube comment analysis for mental health profiling using NLP sentiment patterns.
NLP-based support ticket classification and prioritization using TF-IDF and ensemble ML models.
Medical ML model achieving 98%+ diagnostic accuracy using PyTorch binary classification.
T5-small fine-tuning for abstractive news summarization with ROUGE evaluation metrics.
End-to-end NLP spam detection pipeline for email and SMS using binary classification.
CRAN-published machine learning framework built with a C++ backend and R integration, focused on algorithms, performance optimization, and research.
Professional internship and training
Future Interns
Completed an intensive ML internship building production-grade machine learning systems. Worked on real-world datasets spanning sales forecasting, NLP classification, and AI-powered recruitment tools.
Built an AI-powered resume screening system processing 1000+ resumes using TF-IDF vectorization for candidate-job matching and skill gap analysis.
Developed an NLP-based support ticket classification and prioritization system using TF-IDF and ensemble ML models.
Created a sales and demand forecasting model using Random Forest Regressor on real-world sales data.
Published academic work
This paper presents VectorForgeML, a modular machine learning framework built entirely from scratch in C++ with R bindings. The framework implements 9+ supervised and unsupervised algorithms including Linear Regression, Logistic Regression, Ridge Regression, Softmax Regression, Decision Trees, Random Forests, KNN, PCA, and K-Means Clustering. Key architectural decisions include BLAS/LAPACK integration for hardware-accelerated linear algebra, OpenMP for multi-core parallelism, and zero-copy data exchange between R and C++ via Rcpp.
Professional credentials and coursework
Future Interns
Successfully completed ML internship covering sales forecasting, NLP classification, and AI resume ranking.
CIN: FIT/JAN26/ML4913Completed learning milestone
AI University Curriculum — Completed
Production backend systems with ASP.NET Core, C#, Entity Framework Core, and SQL Server
Alongside AI/ML engineering, I build production backend applications using the .NET ecosystem — including REST APIs, MVC applications, real-time systems with SignalR, and AI-integrated backends. This allows me to ship complete, end-to-end systems rather than isolated model prototypes.
Enterprise security analyzer rebuilt with an ASP.NET Core backend demonstrating how AI integrations (LLM analysis, report generation) layer onto production .NET architecture with Entity Framework Core, SQL Server, React frontend, SignalR real-time updates, JWT authentication, Docker, and GitHub Actions CI/CD.
Full-stack AI chat application built with ASP.NET Core MVC, integrating Groq API for LLM responses, SQL Server for persistent conversation storage, cookie-based authentication, and real-time conversation summarization.
ASP.NET Core Web API with routing, versioning, and middleware
Entity Framework Core, LINQ, SQL Server, code-first migrations
JWT Bearer authentication, cookie auth, role-based access control
SignalR for WebSocket-based live updates and notifications
Connecting LLM APIs, embeddings, and ML models into .NET backends
Docker containerization and GitHub Actions CI/CD pipelines
Verified milestones and recognitions
VectorForgeML — available on the Comprehensive R Archive Network
View on GitHub →VectorForgeML: A Production-Grade Machine Learning Framework in C++
Spanning NLP, computer vision, generative AI, LLMs, RAG, and predictive analytics
View GitHub →Interested in working together? Let's connect.
Pune, Maharashtra, India
musheerayan@gmail.com