Experiment Step 1
Implemented LLM calls producing validated JSON schema outputs.
Projects
Projects across AI, software, and applied problem solving
My work ranges from graduate AI experiments to undergraduate systems projects focused on safety, detection, and practical real-world impact.
GenAI Lab • Nov 2025 - Dec 2025
Review-Driven Product Image Generation
Built a pipeline that converts marketplace reviews into product images using structured LLM extraction and diffusion generation.
Tech Stack
PythonLLMsJSON Schema ValidationDiffusion ModelsPrompt Engineering
Experiment Step 2
Designed modular architecture for ingestion, analysis, prompt construction, and generation.
Experiment Step 3
Added centralized orchestration, logging, and evaluation hooks for deterministic and debuggable runs.
Orchestra AI • Aug 2025 - Dec 2025
AI SEO Experiment
Ran large-scale comparisons between ChatGPT recommendations and Bing rankings to quantify AI-vs-SEO divergence.
Tech Stack
PythonChatGPTBingDataForSEOExperiment Analytics
Experiment Step 1
Executed a 173-case experiment across 9 intent categories.
Experiment Step 2
Automated SERP collection and structured reporting across 40+ keywords using DataForSEO.
Experiment Step 3
Generated findings that informed the roadmap for Spyglasses, an AI-visibility analytics startup.
Undergraduate Project • NIE Mysore • Nov 2021 - Mar 2022
Smart Safety Watch for Elderly People and Pregnant Women
Designed an IoT-enabled safety wearable that combined fall detection, panic alerts, and remote health monitoring support for elderly users and pregnant women.
Tech Stack
FlutterFirebaseArduinoBluetoothGPSSpO2 SensorsIoT
Experiment Step 1
Used sensor-based human-computer interaction and sensor fusion concepts to support fall detection for elderly users in everyday environments.
Experiment Step 2
Included health monitoring signals such as SpO2 and heart rate tracking to support sleep pattern sensing and well-being monitoring for pregnant women.
Experiment Step 3
Built a companion Flutter mobile app to handle heavier processing and send panic alerts with real-time location to designated emergency contacts.
Undergraduate Project • NIE Mysore • Jun 2021 - Jul 2021
Fugitive Detection Traffic System for a Safer Society
Built a Flask-based public safety application that used traffic camera inputs to detect fugitives and weapons and alert law enforcement quickly.
Tech Stack
FlaskPythonCNNComputer VisionPublic Safety Alerts
Experiment Step 1
Applied convolutional neural networks to detect fugitive faces from traffic camera feeds by cross-referencing a law enforcement database.
Experiment Step 2
Added weapon detection to flag public threats and surface alerts as soon as suspicious activity was identified.
Experiment Step 3
Designed the system to notify law enforcement teams through web or mobile interfaces to support faster response.