experience

Built an intelligent vehicle diagnostic assistant analyzing over 4,000 DTCs to streamline fault diagnosis and reduce technician resolution time across vehicle ECUs.

• Implemented traditional RAG pipeline over 4,000+ vehicle DTCs for fault diagnosis and root cause retrieval
• Built a live diagnostic dashboard with real-time vehicle scans via DoIP scanner integration
• Developed AI symptom search mapping natural language inputs to fault codes using SQLite FTS5 and LLM

Re-architected data loading and resolved platform-wide defects to improve performance and reliability for clinic staff and patients.

• Re-architected the data loading strategy on the platform's highest-traffic page, cutting load time by 35%
• Diagnosed and resolved validation and UI defects in the patient registration workflow, updated the underlying schema, and shipped fixes platform-wide
• Partnered with 5+ clients to translate real-world pain points into prioritized engineering fixes across performance, onboarding, and scheduling workflows

Conducted applied AI research in collaboration with the National Institute of Fashion Technology (NIFT) to enhance fabric classification using deep learning.

• Designed and implemented a custom CNN-based model tailored for recognizing culturally rich Indian textile patterns
• Curated and preprocessed a custom dataset of over 2,000 high-resolution fabric texture images spanning diverse weave styles
• Achieved a 91% test accuracy by applying advanced techniques such as regularization, augmentation, and hyperparameter tuning

Analyzed the security posture of over 250 smart home and IoT devices, identifying critical vulnerabilities and misconfigurations.

• Conducted in-depth risk assessments and penetration tests across a variety of device ecosystems
• Proposed and documented actionable mitigation strategies and contributed to detailed technical vulnerability reports
• Supported efforts to improve security awareness in consumer-grade IoT infrastructure through research-driven recommendations