I'm a Computer Science student passionate about building AI systems that solve real problems. My work sits at the intersection of cybersecurity and healthcare AI — two domains where intelligent systems can have a direct, meaningful impact on people's lives. I build end-to-end ML pipelines, deploy production-ready applications, and obsess over clean system design.

AI-powered cybersecurity monitoring and automated incident response platform

Built a real-time threat detection system that ingests and analyzes live log data across distributed systems to identify and respond to security incidents.

• Real-Time Detection: Designed ML-driven anomaly detection on log streams to flag suspicious behavior, attack patterns, and indicators of compromise (IoCs).
• Scalable Backend: Developed a FastAPI-based ingestion pipeline that processes logs from multiple sources (servers, firewalls, cloud services) and normalizes them for analysis.
• Threat Contextualization: Mapped detected events to the MITRE ATT&CK framework, giving immediate insight into attacker tactics and progression.
• Automated Response: Implemented rule-based and threshold-triggered playbooks to isolate systems, revoke credentials, and generate structured incident reports.
• Visualization Layer: Built a React dashboard displaying live alerts, threat severity scoring, and incident timelines for quick decision-making.

Intelligent clinical decision support system for symptom-based diagnosis

Developed an end-to-end machine learning system that transforms patient symptoms and clinical text into ranked diagnostic insights with supporting research.

• Diagnostic Modeling: Trained a multi-label classification model across a wide range of diseases and symptoms to generate probability-ranked differential diagnoses.
• Clinical NLP Pipeline: Used transformer-based models (BERT) to extract structured medical entities from unstructured clinical input.
• Research Integration: Connected to external medical research APIs to retrieve relevant studies, treatment guidelines, and clinical data in real time.
• Automated Reporting: Generated structured outputs including clinical summaries, SOAP notes, and simplified patient explanations using LLMs. Privacy-Focused Design: Built with data anonymization, audit logging, and optional local inference to align with healthcare data standards.

Python, TypeScript, JavaScript, SQL, BashMachine Learning & AIPyTorch, TensorFlow, scikit-learn, Keras, Hugging Face, LangChainNLP & LLMsBERT, GPT-4, RAG pipelines, OpenAI APIBackendFastAPI, Node.js, REST APIs, PostgreSQL, RedisFrontendNext.js 14, React, Tailwind CSS, TypeScriptDevOps & CloudDocker, AWS, GCP, Vercel, GitHub Actions, Pandas, NumPy, Matplotlib, Seaborn, Jupyter

I'm currently deepening my work in large language model fine-tuning, agentic AI systems, and building more robust evaluation frameworks for medical AI models. Always interested in projects that push AI into high-stakes, high-impact domains.

Portfolio: https://anya-rajesh.vercel.app/

GitHub: @anyarajesh1

LinkedIn: http://linkedin.com/in/anya-raj