Aadi Joshi
Building expertise in algorithms, systems, and backend architecture. Research in NLP, neural architectures, RL, and the math of ML—focused on reasoning, robustness, and efficiency in LLMs and multimodal systems. Also active in competitive programming (Codeforces/LeetCode), sharpening problem-solving and algorithmic thinking.
Technical skills
Languages: Python, C++, SQL, Bash, JavaScript
Frameworks: PyTorch, TensorFlow, FastAPI, Flask, Django, React, Node.js
Tools: NumPy, pandas, scikit-learn, OpenCV, spaCy, Sentence Transformers, HF Transformers, FAISS, MediaPipe, Docker, Git, MLflow, W&B
Platforms: Linux, Colab, Jupyter, AWS, GCP, Arduino/ESP32
Education
- Vishwakarma Institute of TechnologyB.Tech Computer Engineering (2024–2028)
Experience
- Software Developer Intern — VIT Pune (Dec 2024–Feb 2025): UI/features for Pariksha Mitra (React/Next)
- Runner-Up — PICT TechFiesta Intl. Hackathon (Feb 2025): Virtual Rehab Assistant; 91% across 30 exercises
- Open Source Contributor — SWoC (Jan–Mar 2025): 10 merged PRs across active repos
Projects
- Android-optimized DR detector using MobileNetV3 + CORAL (15MB TorchScript); 0.83 QWK on 3-fold CV.
- Tech: PyTorch, CORAL, CLAHE, Albumentations. (Jun 2025)
- Personalized physio plan app with real-time motion tracking using MediaPipe and an integrated AI chatbot assistant.
- Tech: React Native, Flask, MongoDB, MediaPipe, Llama. (Jan 2025)
- Academic content manager with role-based access and analytics dashboard.
- Tech: React, Node.js, MongoDB, Supabase, Tailwind CSS. (Mar 2025)