# Atharva Shembade > ML Research Assistant @ USC | MS CS | Agentic AI & LLMOps | Machine Learning | Seeking full-time opportunities since Dec’25 Location: Santa Clara, California, United States Profile: https://flows.cv/atharvashembade ▸ LLM Engineer who ships multi-agent systems, RAG pipelines, and real-time inference infrastructure. ⎯⎯⎯ ▸ Built PaperSage, a production multi-agent RAG platform with self-correcting LangGraph agents and LLMOps observability. ▸ Engineered an end-to-end LLM fine-tuning and inference serving pipeline with quantization and LoRA-based adaptation. ▸ Designed a real-time SSE streaming orchestrator with adaptive backpressure and fault-tolerant stream resumability. ⎯⎯⎯ ▸ Previously: ML Research Assistant at USC (computer vision on aerospace materials) and ML Intern at Renishaw (defect detection for industrial manufacturing). ⎯⎯⎯ ▸ M.S. Computer Science - University of Southern California ▸ github.com/AtharvaShembade | Open to LLM Engineer / AI Engineer roles ## Work Experience ### Machine Learning Research Assistant @ University of Southern California Jan 2024 – Jan 2025 | Los Angeles, CA • Neutralized manual inspection bottlenecks for 100 GB of micro-CT aerospace data by engineering a void morphology pipeline, accelerating quality assurance cycles by 95% • Mitigated safety-critical failure risks in aerospace materials by synthesizing a hybrid 3D CNN-Swin Transformer, surpassing legacy benchmarks by 18% with a 0.92 Mean IoU • Rectified data integrity anomalies for irregular geometries via hierarchical attention mechanisms, driving a 15% gain in segmentation precision over baseline models ### Research Intern @ Renishaw India Jan 2023 – Jan 2023 | Pune, Maharashtra, India • Formulated Capsule Networks to handle spatial hierarchies, mitigating industrial data scarcity and delivering a 10% accuracy lift over CNNs while reducing labeled data requirements by 50% • Established a production pipeline using ONNX and ANOVA, compressing the model evaluation cycle by 40% and eliminating false-positive architecture candidates prior to deploymen ### Machine Learning Intern @ Renishaw India Jan 2022 – Jan 2023 | Pune, Maharashtra, India • Scaled autonomous quality control to handle 10,000 daily inspections, achieving defect categorization with 99.9% accuracy, by fusing Faster R-CNN precision with YOLO inference speeds • Recovered 10 hours/week and prevented assembly line stoppages by engineering a sub-50ms stacked ensemble pipeline, reducing false positives by 4x while handling 5 GB/hour throughput • Optimized GPU memory allocation during migration from TensorFlow to PyTorch, shortening model training cycles by 3 hours and accelerating R&D experimentation ## Education ### Master of Science - MS in Computer Science University of Southern California ### Bachelor of Technology - BTech in Electronics and Telecommunication Vishwakarma Institute Of Technology ### Loyola High School & Junior College Pune ## Contact & Social - LinkedIn: https://linkedin.com/in/atharva-shembade - Portfolio: https://papersage-research.vercel.app/ - GitHub: https://github.com/AtharvaShembade - Portfolio: https://papersage-research.vercel.app --- Source: https://flows.cv/atharvashembade JSON Resume: https://flows.cv/atharvashembade/resume.json Last updated: 2026-04-18