I build AI systems that listen, plan, and act.

AI/ML Engineer Intern, Full-Stack

Architected and deployed a pediatric speech therapy platform that runs clinical assessments in 8 Indian languages.

• Designed and implemented a Q-learning adaptive difficulty controller that tunes each child's practice games from rolling pronunciation accuracy and response latency.
• Built a real-time pronunciation scoring engine chaining Faster-Whisper, Vosk and Kokoro/Coqui TTS with phoneme-level acoustic checks (Librosa, Epitran), with sub-300ms feedback latency across submodules (BreathQuest, VaakMirror, Chime, VoiceHurdleRace).
• Engineered a LangGraph pipeline that runs diagnostic tests, classifies errors by place, manner and voicing, and writes structured clinician reports.
• Added MediaPipe facial landmark tracking for live oral-motor exercise feedback.
• Owned full-stack delivery: FastAPI, PostgreSQL, COPPA-compliant parental consent, therapist and parent dashboards, React and Vite.

AI/ML Engineer Intern

Built audio ingestion, speech-to-text and RAG pipelines on Azure for a wealth-management platform serving 10,000+ mutual fund distributors.

• Tuned speech-to-text (Sarvam AI, Whisper) with spectral noise reduction for mixed Indian English and Hindi advisor calls.
• Built a RAG summarization layer that turns advisory calls and portfolio reviews into structured CRM action items.
• Automated PDF, DOCX and transcript extraction into vector stores, cutting manual documentation turnaround by 80%.

B.Tech, Robotics and Artificial Intelligence

Shanmugha Arts, Science, Technology and Research Academy.

Certifications

• SWAYAM, IIT Madras: Comprehensive Electronics and Embedded Systems (in-person practical program)
• IIoT Certification, Ministry of Heavy Industries program (project-based evaluation)

Languages

English, Hindi, Tamil