ECE Final-Year @ BMS College of Engineering, Bengaluru
🎯 Focused on AI-driven Signal Processing, Communication Systems & Intelligent DSP
I’m fascinated by how information travels — through air, wires, or neurons — and how intelligence can adapt to that noise.
My work blends machine learning, signal processing, and system design, building end-to-end frameworks that learn from and react to dynamic signals.
I’m currently aligning my research and engineering work toward and aiming to specialize in AI for Communication & Signal Systems.
| Project | Domain | Current Focus |
|---|---|---|
| 🛰 SkyTune | AI for SATCOM | Learning-based ModCod adaptation under fading conditions |
| 🧩 SigFlow | Software Signal Pipeline (Python + C++20) | Integrating DSP ↔ ML for adaptive signal chains |
| 🧠 Neural Transceiver | Deep Comms Learning | Autoencoder-based encoder-decoder over AWGN/Rayleigh channels |
| 🔍 Neural Channel Estimation | OFDM + CNN | CNNs for channel-tap inference & BER optimization |
| 📡 Anomaly Detection in Sensor Streams | Time-series ML | LSTM Autoencoder for industrial signal anomalies |
- Signal Intelligence: adaptive filtering, FFT/OFDM, Rayleigh/AWGN modeling
- Learning Systems: CNNs, LSTMs, Transformers, Autoencoders, RL agents
- Integration: Python ↔ C++20 (PyBind11), MATLAB/CommPy simulations
- Applied Math: Probability, Linear Algebra, Optimization, Information Theory
Core Libraries: NumPy · SciPy · scikit-learn · FFTW · Eigen · PyBind11 · TensorBoard
- ⚙️ SigFlow reduced DSP latency by ~30 % using PyBind11 C++20 core
- 🛰 SkyTune prototype improved throughput by ~18 % at equal BER
- 🧠 EEG Classification achieved 92 % F1-score using spectral features
- 🔊 Adaptive LMS Canceller delivered ~20 dB SNR gain on live audio
- 🎓 Top 5 % nationwide — NPTEL Machine Learning (IIT-Madras)
- AI-based Communication Systems
- Neural Channel Estimation & Modulation Recognition
- Adaptive DSP and RL-driven PHY optimization
- Biomedical & Acoustic Signal Intelligence
| Phase | Timeline | Focus |
|---|---|---|
| Phase 1 – Integration | Nov – Dec 2025 | Build & benchmark SigFlow v1.0 (Python + C++20) |
| Phase 2 – Domain Depth | Jan – Feb 2026 | Paper + Experiment sprint on AI-Comms / Signal AI |
| -- Rest will be updated later |
- 🎸 Amateur music producer — experiments with synth-based sound design
- 🚲 Love a good long ride
- 🌊 An avid swimmer
“Build systems that understand their own noise.”
— Dibyojyoti Bhattacharjee