Reinforcement Learning for Antenna Design: A Deep Dive
Exploring how uncertainty-aware RL agents can optimize electromagnetic structures, reducing simulation costs by 65% through hybrid surrogate modeling approaches.
Read More βHello, I'm
AI/ML Engineer
Turning cutting-edge research into real-world AI systems, with hands-on experience in computer vision, NLP, and reinforcement learning strengthened by published work and practical MLOps deployment.
AI/ML Engineer specializing in transforming advanced research into scalable, production-ready systems. I engineered a reinforcement learning framework that delivered a 65% reduction in EM simulation cost by optimizing antenna design with uncertainty-aware SAC agents and hybrid surrogate models.
My experience spans multilingual NLP pipelines, computer vision systems, large-scale RAG architectures, and secure communication platforms. I combine deep learning expertise with strong MLOps capabilitiesβdeploying models using FastAPI, Docker, Vertex AI, and optimized GPU workflows.
Based in Pithapuram, Andhra Pradesh, India, I focus on building AI solutions that are measurable, reliable, and impactful. My published research and applied engineering work reflect a commitment to creating real-world AI systems that solve complex technical and business challenges.
Architected reinforcement learning framework automating antenna geometry design, reducing CST simulation calls by 65%. Implemented uncertainty-aware Soft Actor-Critic agent with LightGBM + MLP surrogate models, custom Gymnasium environment, and fabrication-feasible reward shaping.
Built production-ready RAG pipeline supporting 12+ Indian languages with multilingual dense embeddings and cross-encoder reranking. Integrated OCR for scanned PDFs, semantic chunking for long documents, and deployed FastAPI service.
Engineered CLI-based encrypted messaging system using AES-256 + RSA-2048 hybrid cryptography. Implemented secure key exchange protocol, encrypted TCP sockets with SHA-256 integrity verification, and multi-user real-time messaging with zero-knowledge architecture.
Developed CNN achieving 89% test accuracy across 10 land-use classes on satellite imagery. Integrated Grad-CAM explainability for feature attribution and model interpretability.
Google (via EduSkills)
Palo Alto Networks (via EduSkills)
Altair (via EduSkills)
Bapatla Engineering College, Andhra Pradesh
Aditya College of Engineering, Surampalem, AP
Ravindra Bharathi High School, Pithapuram, AP
SQL β’ R Programming β’ Data Cleaning β’ Tableau β’ Business Communication
March 2025Network Security β’ Python β’ Linux β’ Cloud Security β’ Network Architecture
May 2024Prompt Engineering for GenAI Workflows
May 2025AI Strategy β’ ML Fundamentals β’ Business Use-Cases
February 2025TensorFlow β’ Image Classification β’ Object Detection β’ Product Search
2024-2025Engineering Architecture β’ Virtualization Concepts
2024TCP/IP β’ Routing β’ DNS β’ Network Performance
2024Application No.: 202541014595 A | Published: March 7, 2025 (India)
Presented at ICMOCE 2025, IIT Bhubaneswar
Presented at 16th ICCCNT 2025, IIT Indore
Exploring how uncertainty-aware RL agents can optimize electromagnetic structures, reducing simulation costs by 65% through hybrid surrogate modeling approaches.
Read More βA practical guide to implementing production-ready retrieval-augmented generation pipelines for Indian languages with semantic chunking and cross-encoder reranking.
Read More βLearn how to implement hybrid AES-256 and RSA-2048 cryptography with secure key exchange protocols for zero-knowledge architecture messaging.
Read More βI'm always interested in hearing about new opportunities, collaborations, or just having a conversation about AI/ML. Feel free to reach out!