Jul 2024 – Jul 2025 · 1 yr

• Worked on performance optimization for AI/ML frameworks and numerical libraries targeting Fujitsu’s next-generation MONAKA ARM-SVE supercomputing 2nm 288-core processor (successor to Fugaku).
• Key Contributions & Responsibilities:
• Worked on Llama.cpp Scaling issues for ARM CPU (Q4)
• Optimized OpenBLAS Inner kernels and enabled OpenMP Task based thread composability, improving scalability on many-core systems.
• Designed and implemented Threading-as-a-Service SDK (Layer) to manage parallel scheduling, resource allocation and execution efficiency for ML DL Frameworks.
• Developed and contributed ARM-SVE intrinsics and vectorized compute kernels for AI/ML workloads.
• Optimized oneDAL (oneAPI) for ARM-SVE, achieving significant improvements in ML & analytics algorithms with lower compute cost Contributed to OSS.
• Worked with time-series pipelines, PyTorch, StateModels, and Numba JIT optimization techniques & Frameworks.
• Performed profiling, performance tuning, memory optimization, and cross-architecture benchmarking (ARM vs x86).
• Contributed patches and discussions to open-source communities and ecosystem collaborations.
• Achievements:
• Filed innovation/IP contributions and received internal Innovation Award.
• Spot and Q awards from Fujitsu
• Work referenced in technical blogs, articles, and foundation working groups.
• Visited Fujitsu Advanced Computing Unit (Japan) for technical collaboration, architecture discussions, and performance demo presentations.

Jul 2023 – Jul 2024 · 1 yr

Mar 2022 – Jun 2023 · 1 yr 3 mos · Bangalore Urban, Karnataka, India · On-site

• Worked on IMX500, the world’s first vision sensor with on-sensor AI processing.
• Optimized and quantized CNN models for real-time edge inference on the IMX500 DSP.
• Ported models from PyTorch/TFLite to the device pipeline and benchmarked performance.
• Developed backend services for model deployment, job monitoring, and device management.
• Contributed to CI/CD workflows and cloud deployments using GitHub Actions, Jenkins, and AKS.

Jan 2022 – May 2023 · 1 yr 4 mos · New York, United States · Remote

• Ideation and development and Onboarding
• Directed product architecture and technical roadmap for FINTEC-AI-based enterprise solutions.
• Managed cross-functional engineering teams, client discussions, and solution delivery cycles.

Sep 2019 – Feb 2022 · 2 yrs 5 mos · Bangalore · On-site

• Developed and deployed end-to-end Computer Vision, OCR/IDP, and Edge-AI solutions for enterprise clients, covering the full pipeline from data to production deployment.
• Designed AI pipelines for object detection, classification, segmentation, pose estimation, and OCR/ICR/NER across multiple domains.
• Built and optimized data workflows including dataset creation, synthetic data generation, annotation automation, and augmentation strategies.
• Delivered POCs, MVPs, and production-ready systems by collaborating closely with customers and pre-sales teams.
• Led feature development, model benchmarking, deployment on cloud and edge devices, and post-deployment performance tuning.
• Mentored and led a team of 15+ interns and engineers in deep learning workflows, optimization techniques, and deployment practices.
• Selected Projects:
• TXTR (OCR / ICR / IDP Platform): End-to-end pipeline for document understanding (handwritten + printed) used across KYC, invoices, resumes, banking & financial docs.
• Warehouse.AI: Automated pallet & box counting, barcode reading, and product validation for cold-storage logistics.
• OutOfDanger: Workplace safety analytics (PPE detection, mask detection, distancing, occupancy & attendance) over CCTV/IP camera streams.
• Face Recognition System: Real-time recognition pipeline with registration & identity management, optimized for mask-enabled and long-distance scenarios.
• PhotoShield: Screen privacy monitoring system to detect external recording devices (DSLR / mobile / spy cams)
• Architect Plan Parsing: Parsed blueprint PDFs to detect new cable additions and extracted lengths/types for RPA pipeline automation.

Jan 2018 – Jan 2019 · 1 yr · Dhanbad Area, India · On-site

• Assisted undergraduate computer science students in their course on algorithms.
• Teaching programming skills and techniques in Data Structure, Advance Algorithms in C/C++, data mining, Machine learning, Operating system to a class of 180+ second year, third-year and freshmen. Responsible for teaching and performance evaluation in the lab, designing assignments, and examination questions.