Siddharth Medhamurthy
AI/ML Engineer exploring Computer Vision and Model Explainability
Bengaluru, Karnataka
Ramaiah Institute of Technology
B.E. in Computer Science (AI & ML)
Expected Graduation: June 2026
Relevant Coursework:
Machine Learning, OS, DBMS, Big Data, Networks
Professional Experience
- Created AI optimization tool leveraging evolutionary algorithms to fine-tune 20 camera ISP parameters
- Increased object detection accuracy by 15% using evolution model-assisted tuning methods
- Designed cross-platform GUI with PySide6, improving usability and streamlining workflows for engineers
- Reduced Time-to-Market for Camera products by 50%
- Accelerated camera system prototyping by reducing manual tuning cycles by 90%
- Extended AI optimization capabilities for advanced camera systems
- Collaborated on next-generation ISP parameter optimization algorithms
- Enhanced GUI functionality and user experience improvements
- Optimized ARM assembly code to reduce computations by 50% on edge devices
- Engineered AI pest trap using image classification, reducing product cost by 85% to under $45
- Deployed software for remote monitoring and firmware updates for multiple devices via LTE/3G IoT modules
- Improved device performance and reliability through low-level embedded programming
Selected Projects
ML ISP Tuning Suite
Jun 2024 – Present
Developed performance suite processing 100 images/batch in under 10 minutes, making manual analysis redundant. Achieved 15% higher object detection accuracy over industry benchmarks, enabling pilot collaborations with STMicroelectronics and Renesas.
Uterine Fibroids Segmentation
Jun 2024 – Present
Trained UNet++ model with EfficientNet B5 Encoder on 1600+ augmented ultrasound images. Achieved 85% IOU score, beating all other research efforts. Secured ₹4 Lakh funding from Ramaiah's venture. Collaborating with Ramaiah Hospital for larger Indian dataset.
YouTube Creator AD-Intelligence Platform
Mar 2025 – Present
Built web application providing creators AI tools for real-time comment analysis and engagement. Features Comment Summarizer, AI Chat via Gemini integration, and insights dashboard with region-wise analytics and engagement metrics.
Lekha - Legal Productivity Platform
2024
Developed Lekha, a role-based legal productivity app tailored for Indian lawyers and law interns. The platform offers distinct onboarding and dashboards for Senior Advocates, Junior Advocates, and Interns, with features like case management, court diary integration, intern supervision, and document organization. Interns can request access to advocates, submit resumes, and assist with real-time case tracking and research. The app includes AI-powered features such as voice-based case status queries, case summarization via PDF parsing, and smart diary logging during or after court proceedings. Despite filing writ petitions to the High Court of Karnataka, we were denied access to necessary court data APIs by the e-Committee, stalling deeper integration with high court databases and live cause lists. Until government approval is granted, the product remains in limited release.
Smart Pesticide Deployment ML System
Jan 2024 – May 2024
Architected full-stack precision agriculture application deployed on research farms at Western Sydney University and IIHR. Developed custom image preprocessing enhancing crop image quality by 30%. Applied object detection models achieving 95% precision for targeted pesticide deployment, reducing overuse by 30%.
Let's Connect & Collaborate
Ready to discuss AI/ML opportunities, research collaborations, or innovative projects? Fill out the form below to unlock my contact details!
Tell Me About Your Project or Opportunity
Contact Information
Open to internships, research collaborations, and full-time opportunities in AI and Computer Vision.