AI/ML Intern at MRIKAL (2026-01 – 2026-04)
- Developed an NLP-powered chatbot extension for Apache Superset and Preset dashboards to enable natural language querying.
- Improved query accuracy to 90% on simple dashboards and 82% on complex dashboard workflows through LLM evaluation.
- Implemented a real-time SQL generation pipeline to convert user prompts into optimized database queries.
- Created an LLM-based intent classification system to route informational and analytical queries efficiently.
ML Engineer Specialist – AI Trainer at Invisible Technologies (2025-10 – 2025-12)
- Evaluated and optimized 90+ machine learning problem-solving workflows for Meta's ML Engineering project.
- Compared multiple implementation approaches based on feasibility, scalability, and reasoning accuracy.
- Analysed edge cases and workflow constraints to improve consistency and decision quality in AI-generated solutions.
Data Scientist Intern at WESEE (Indian Navy) (2025-07 – 2025-09)
- Developed a YOLOv11n-based object detection system to identify aircraft, submarines, and ships from defense imagery datasets.
- Processed and annotated a 70GB dataset using LabelMe and achieved 87% detection accuracy after model optimization.
- Engineered an offline OCR + LLM document understanding pipeline with 94% extraction accuracy from scanned PDFs.
- Integrated Mistral-7B and semantic similarity search to enable document-grounded Q&A, reducing response time by 40%.
Robotics Intern at MESHLINTECH (2025-04 – 2025-07)
- Created 3D CAD models and simulations for an automated garland-making machine using Fusion 360 and Blender.
- Optimized the knot-making mechanism, increasing operational throughput by 25%.
- Improved machine automation efficiency by 40%, reducing manual intervention and operational costs.
