🌐 Portfolio · 💼 LinkedIn · 🐙 GitHub · ⚡ LeetCode · 🤗 HuggingFace · 🔗 Social Hub
I am a Full-Stack Developer and Machine Learning Engineer focused on building software across web, mobile, and cloud ecosystems. My work spans building real-time AI conversation systems, training transformers from scratch, implementing medical imaging pipelines, and developing cross-platform applications.
I come from a Mechatronics background (diploma) and transitioned into software — bringing a unique blend of hardware thinking and engineering discipline to every project.
- Vibeathon 2025 Winner – 1st place among 367 teams at India's AI hackathon by Replit × Polaris
- Patent Filed – Infant Cry Classification System (Application No: 202541047092)
- Finalist – Smart India Hackathon 2025
- 4th Place – TiE Global Summit Ideathon 2024
- Open Source Contributor – building modular tooling ecosystems
- B.Tech Candidate – Computer Science & Engineering @ Dayananda Sagar University (2022–2026)
Cross-platform open-source music streaming ecosystem
Native Android (Kotlin/Compose), iOS (SwiftUI), Web (React), and Backend API with background playback, local storage management, Cashfree billing integration, and centralized media server architecture. Smart search engine with auto-complete and fuzzy indexing. Background downloader with storage optimization and gapless playback.
Real-time AI conversation system with multi-persona orchestration
A modular application pipeline orchestrating real-time communication systems, voice synthesizers (TTS/STT), dynamic memory modules, and specialized AI personas. Features a JSON-configured persona framework with contextual injection, low-latency audio processing pipeline, and a responsive Electron desktop shell.
🔗 Repositories · Website
AI-assisted developer workflow automation — Vibeathon 2025 Winner
Automation utility helping developers streamline complex code compilation, optimization, and system integration using LLM orchestration. Designed and deployed within a 36-hour coding window. Features intelligent context generation from local directories and real-time execution tracking.
🔗 GitHub
🧠 OwnLLM — Transformer from Scratch (Python, PyTorch, CUDA)
Complete custom implementation of a Decoder-only Transformer model with multi-head self-attention, Rotary Positional Embeddings (RoPE), and a custom O(n) Byte-Pair Encoding (BPE) tokenizer. Mixed-precision FP16 training with active loss scaling.
🔗 GitHub
🔬 CryDecoder — Infant Cry Classifier (Python, TensorFlow, Flask)
Deep learning CNN-LSTM architecture trained on acoustic cry patterns achieving 91%+ validation accuracy. Patent-filed system with symptom-to-disease mapping covering 4,500+ symptoms to 400+ medical entities. Deployed with user feedback portal.
⚡ Concurrent API with Load Balancing (Go, Python, PostgreSQL, Redis)
High-concurrency REST engine comparing processing pipelines across Go, Node.js, and Python. Reached 5,300 RPS under stress testing. Implements custom rate limiters, LIFO/FIFO queue schemas, and task schedulers.
🔗 GitHub
🏥 SMIRE-AI — Medical Imaging System (Python, TensorFlow, pydicom)
Semantic segmentation of MRI/CT scans using U-Net architectures with skip connections. Elastic deformation augmentation, DICOM parsing, and PDF report generation.
🔗 GitHub
👁️ OpenGUI — AI-Powered Desktop Automation (Python, PyTorch, OpenCV)
Fine-tuned YOLOv8 to detect UI elements (buttons, fields, bars) and orchestrate cross-platform mouse/keyboard automation. Automates repetitive workflows through visual parsing.
🔗 GitHub
🌐 CureConnect — AI Disease Prediction System (Python, Flask, Scikit-learn)
Clinical prediction platform using voting ensemble models (SVM, Random Forest, Gradient Boosting). Maps 4,500+ symptom profiles to diagnostic outputs with SpaCy NLP pipeline.
🔗 GitHub
🖥️ Infinity-CLI — Smart Terminal Interface (Node.js, TypeScript)
Natural-language CLI that generates executable shell scripts from plain English. Fuzzy lookup, contextual history tracking, and error recovery parsing.
🔗 GitHub
🤝 HumanLoop — AI Interview Simulator (React, Node.js, Socket.io)
Interactive interview platform with real-time evaluation reports. Voice synthesis with adaptive prompt structures and multi-attribute scoring system.
🔗 GitHub
📱 Daly Streak — Habit Tracking App (Flutter, Firebase)
Offline-first habit tracker with custom FCM notifications and optimized Firestore sync layers supporting fully offline use.
🔗 GitHub
🎥 AI Video Interviewer — Emotion Tracking Portal (Node.js, WebRTC, Chart.js)
Real-time evaluation platform using Face++ APIs for engagement scoring, pose estimation, and expression analysis via WebRTC streaming.
🔗 GitHub
💰 SocialFlo — Content Automation & Subscriptions (React, Node.js, MongoDB)
Multi-platform content scheduler with Stripe payment integration, automated webhooks, and Redis/Bull queue scheduling.
🔗 GitHub
| Project | Link | Status |
|---|---|---|
| Personal Portfolio | personal-dashboard-h4t1.onrender.com | Active |
| DevFlow.Ai | devflow-ai-ztwh.onrender.com | Active |
| CST Notes | cst-notes-71801.web.app | Under Maintenance |
| NexHire AI Interviews | nexhire.pythonanywhere.com | Under Maintenance |
| CryDecoder AI | crydecoder.web.app | Active |
| Fluxion Engine | fluxionengine.app | Active |
| Profile Hub | beacons.ai/Satyam_Pote | Active |
SocialFlo (Germany) · Dec 2024 – Jun 2025 · Remote
Full-stack training engine (React.js, Node.js, Express.js, MongoDB). Multi-platform scheduling pipelines for social media with media adjustments. JWT authentication, RBAC, and Swagger/OpenAPI documentation.
Dayananda Sagar University, Bengaluru · 2022 – 2026 · CGPA: 7.8/10
Core: Algorithms, Systems Programming, Databases, Computer Networks, Machine Learning.
NTTF, Belgaum · 2020 – 2023
Smart Factory Systems, automation engineering, PLC, microcontrollers.
🔗 All Links · 📧 satyampote9999@gmail.com · 📍 Bengaluru, India