Computer Vision & Deep Learning Enthusiast

Python Developer focused on backend systems and computer vision, with a solid foundation in AI and deep learning theory. Experienced in building automation pipelines and real-world applications using FastAPI, OpenCV, and MediaPipe. Currently developing projects with PyTorch and TensorFlow to translate theoretical ML concepts into practical, scalable solutions.

Technical Skills

• Languages & Scripting: Python, Bash
• AI/ML & Computer Vision: FastAPI, scikit-learn, OpenCV, MediaPipe, PyTorch, TensorFlow
• Data Handling & Visualization: pandas, NumPy, Matplotlib, Seaborn
• Web Development: NextJS, ReactJS, TailwindCSS, HTML, CSS

Work Experience

OpenAstronomy — Contributor @ SunPy Project (Jan 2025 – Feb 2025)

• Resolved a critical bug in SunPy’s SRSClient by replacing outdated NOAA FTP endpoints with HTTPS, restoring ~100% CI/CD pass rate across 30+ builds per week.
• Coordinated directly with NOAA API maintainers to verify updated links, reducing future API breakage risk by ~80%.
• Improved data reliability for 1,000+ monthly users and downstream scientific tools by merging fixes into SunPy’s core codebase.
• Fixed UI layout issues on OpenAstronomy.org, enhancing footer responsiveness, theme consistency, and visual flow across ~10 pages.

Education

• B.Tech, Computer Science

  Rishihood University (August 2028)

• High School Diploma, CS

  National English School (March 2024)

Projects

whispercast (March 2025 - April 2025)

GitHub | Website

• Built an AI-powered CLI tool for generating podcasts and audiobooks using Python, reducing manual content prep time by ~80%.
• Automated script creation via OpenAI LLMs and text-to-speech conversion using Coqui-TTS, enabling end-to-end audio generation in under 15 seconds.
• Integrated Wikipedia, Google News, and Reddit APIs into the AI podcasting tool, delivering real-time, topic-specific content updates for each audio segment and improving content freshness by 99%.
• Designed an interactive learning mode that let users query PDFs and Docs, increasing engagement by approx. 2x in test runs.
• Reduced file clutter by auto-managing playback files and temporary content, keeping storage usage under ~50MB per session.

melody.cli (May 2024 - Jun 2024)

GitHub | Website

• Created a terminal-based music player with ~0.5s average latency from command to playback.
• Integrated a smart caching system that reused previous downloads, cutting network usage by up to 60% across sessions.
• Designed for devs and low-resource systems, keeping RAM usage below 100MB even during continuous playback.
• Tested with 50+ tracks across genres to validate consistency and fault tolerance of autoplay queue logic.

meta-input (Feb 2023 - May 2023)

GitHub

• Built a gesture-based controller using OpenCV + MediaPipe that replaced traditional inputs with hand gestures, achieving ~90% gesture recognition accuracy.
• Enabled full mouse control, volume/brightness adjustment, and air typing with minimal latency (~1s response on average hardware).
• Tuned system to maintain stable tracking at 30 FPS, even in medium-light conditions, improving UX consistency.
• Reduced user reliance on peripherals by offering hands-free operation for basic OS tasks — tested successfully over 20+ live sessions.