Open to opportunities · Software & AI Engineering

Anya Rajesh

Building and deploying full-stack AI solutions that integrate real-time data pipelines, retrieval-augmented generation, and privacy-first architectures. Focused on creating impactful, accessible applications that solve real problems.

View Projects
2+
Projects Built
10+
AI Models Trained
GitHub Stars
Project 01

Vigil AI

Intelligent Cybersecurity Monitoring System

View on GitHub →

What is Vigil AI?

Vigil AI is an open-source, AI-native cybersecurity monitoring platform designed to help organizations detect threats before they become incidents. It combines machine learning anomaly detection, real-time log analysis, and intelligent alerting into a unified security operations toolkit.

Built on a FastAPI backend with React frontend, Vigil processes raw log streams from any source — web servers, firewalls, cloud platforms — and applies trained models to surface anomalies, suspicious patterns, and potential intrusions in real-time.

# vigil threat log
[✓] 192.168.1.1 → SAFE (conf: 0.98)
[✓] 10.0.0.5 → SAFE (conf: 0.95)
[!] 203.0.113.42 → WARN (conf: 0.71)
[⚠] 198.51.100.2 → THREAT(conf: 0.94)
→ MITRE: T1078 Valid Accounts
→ Action: Alert triggered
Processed: 14,821 events/sec

Core Capabilities

🛡️

Real-Time Threat Detection

Continuously monitors system logs, network traffic, and user behavior using ML anomaly detection algorithms to identify threats the moment they emerge.

🧠

AI-Powered Behavioral Analysis

Trained on thousands of attack patterns, Vigil AI understands the difference between normal activity and subtle indicators of compromise (IoCs).

📊

Intelligent Dashboard

Visual threat maps, risk scoring, incident timelines, and alert prioritization — all in a clean, operator-friendly interface built for speed.

Automated Incident Response

When a threat crosses confidence thresholds, Vigil can automatically trigger response playbooks: isolating hosts, revoking credentials, and generating reports.

🔍

MITRE ATT&CK Mapping

Every detected event is mapped to the MITRE ATT&CK framework, giving security teams immediate context on attacker tactics, techniques, and procedures.

📁

Multi-Source Log Ingestion

Ingests from web servers, firewalls, SIEM systems, cloud providers, and custom endpoints via a unified Elasticsearch-backed pipeline.

Tech Stack

PythonFastAPIReactElasticsearchPyTorchOpenAI APISQLite/PostgreSQLDockerGrok ParsingMITRE ATT&CK
Project 02

MedInsight AI

Intelligent Medical Diagnostics & Health Analytics Platform

View on GitHub →

What is MedInsight AI?

MedInsight AI is an intelligent medical analytics platform that bridges the gap between raw patient data and actionable clinical insight. Designed for healthcare professionals, students, and researchers, it uses advanced NLP and machine learning models to assist with disease identification, health risk assessment, and medical literature discovery.

The platform's symptom-driven diagnostic engine processes hundreds of clinical indicators simultaneously, providing ranked differentials with confidence scores and relevant medical context — reducing diagnostic guesswork and supporting faster, better-informed clinical decisions.

# Differential Diagnosis Output
Symptoms: fatigue, frequent urination, blurred vision, elevated BP
Hypertension87%
Type 2 Diabetes74%
Anemia61%
→ Recommend: HbA1c, CBC, BMP panel

Core Capabilities

🩺

Symptom-Based Disease Prediction

Select symptoms from a comprehensive library and receive AI-powered differential diagnoses ranked by probability, with supporting clinical context.

📈

Patient Data Analytics

Upload patient vitals, lab results, or historical records and receive trend analysis, outlier detection, and risk stratification powered by ML models.

🔬

Medical Literature Synthesis

Integrates with PubMed and Semantic Scholar to surface relevant research papers, clinical trials, and treatment guidelines for any queried condition.

💊

Drug Interaction Checker

Cross-references prescribed medications against a drug interaction database, flagging contraindications and adverse combination risks automatically.

📋

Clinical Report Generation

Automatically generates structured clinical summaries, SOAP notes, and patient-friendly plain-language explanations from raw diagnostic data.

🛡️

Privacy-First Architecture

HIPAA-aligned design with local inference options, data anonymization pipelines, and audit logging to protect sensitive patient information.

200+
Symptoms Indexed
Comprehensive symptom library
50+
Conditions Diagnosed
Across major disease categories
94%
Model Accuracy
On validation dataset

Tech Stack

PythonStreamlitTensorFlowscikit-learnNLP (BERT)PubMed APIPandasNumPyFastAPIOpenAI GPT-4
Toolkit

Skills & Technologies

Languages
PythonTypeScriptJavaScriptSQLBashR
ML / AI Frameworks
PyTorchTensorFlowscikit-learnKerasHugging FaceLangChain
Data & Analytics
PandasNumPyMatplotlibSeabornPlotlyJupyter
Backend & APIs
FastAPINode.jsREST APIsGraphQLPostgreSQLRedis
Frontend
ReactNext.jsTailwind CSSTypeScriptFramer Motion
DevOps & Cloud
DockerAWSGCPVercelGitHub ActionsElasticsearch