Software Developer – Advanced Micro Devices (AMD), AI Team
I am a software developer at AMD’s AI Team, passionate about building robust, high-performance systems for artificial intelligence and high-performance computing (HPC). My work focuses on optimizing AI frameworks, accelerating model inference, and pushing the limits of compute efficiency.
Beyond my professional role, I am deeply committed to open-source collaboration and knowledge sharing. I regularly contribute to major projects, mentor other developers, and actively participate in the global tech community.
- Programming Languages: C++, Python, Java
- AI & ML Frameworks: TensorFlow, ONNX Runtime, PyTorch, vLLM
- Systems & Performance: Kernel optimization, DevOps, scalable AI infrastructure
- Special Interests: Robotics, HPC, LLM serving, teaching
-
TensorFlow Java:
Enabled pluggable device support for greater flexibility in deployment.
View Commit -
AMD ZenDNN TensorFlow Plugin:
Achieved a 30% performance boost for recommendation system models by optimizing computational kernels.
View Commit
My research interests span deep learning, fuzzy logic, climate science, and advanced streamflow modeling for climate change adaptation.
Deep learning algorithms and their fuzzy extensions for streamflow prediction in climate change framework
Authors: Rishith Kumar Vogeti, Rahul Jauhari, Bhavesh Rahul Mishra, K Srinivasa Raju, D Nagesh Kumar
Journal: Journal of Water and Climate Change, 2024, Vol. 15, Issue 2, pp. 832-848 (IWA Publishing)
Summary:
This study analyzes the capability of advanced deep learning methods—including CNN, LSTM, CNN-LSTM, and their fuzzy extensions—to predict streamflow in India’s Lower Godavari Basin. Fuzzy-based models, especially fuzzy CNN-LSTM, demonstrated significant accuracy improvements over classical algorithms. Projections for four different socioeconomic pathways (SSPs) reveal nuanced shifts in streamflow, offering actionable insights for climate policy and adaptive decision-making. Statistical tests like Mann–Kendall and Pettitt were used to validate trends and change points, highlighting the power of fuzzy deep learning in environmental modeling.
Boosting algorithms for projecting streamflow in the Lower Godavari Basin for different climate change scenarios
Authors: Bhavesh Rahul Mishra, Rishith Kumar Vogeti, Rahul Jauhari, K Srinivasa Raju, D Nagesh Kumar
Journal: Water Science & Technology, 2024, Vol. 89, Issue 3, pp. 613-634 (IWA Publishing)
Summary:
This work investigates the effectiveness of five boosting algorithms—AdaBoost, CatBoost, LightGBM, NGBoost, and XGBoost—for simulating and projecting streamflow under multiple climate change scenarios. Using decades of climate and hydrological data, the study found that all algorithms performed well (high KGE values), with NGBoost excelling in future projections. Ensemble approaches were also explored, underscoring the potential of boosting methods for robust climate-resilient water resource planning.
I am passionate about:
- Advancing the boundaries of AI and high-performance computing
- Open source as a catalyst for innovation and collaboration
- Creating and serving large-scale ML/LLM systems efficiently
- Robotics and applied artificial intelligence
- Teaching, mentoring, and making technical concepts accessible
Curious about my work?
Check out my pinned repositories and contributions below, or reach out to start a conversation about AI, HPC, open source, or anything tech!


