Senior Research Specialist [AI Systems Architect | Edge & Autonomous Systems]

• Nokia, Munich, Germany
• Duration: January 2023 - Present
• Description: Systems Architect specializing in high-performance AI infrastructure and hardware-software co-design. Expert in bridging the gap between frontier ML models and custom GPU runtimes, memory-efficient scheduling, and kernel-level optimizations. Proven track record of eliminating system-level bottlenecks (PCIe, VRAM fragmentation, communication latency) for large-scale training and inference. Passionate about architecting the next generation of scalable AI systems.
• Core Responsibilities:
  • Architected high-performance AI software platforms for distributed edge environments, focusing on minimizing scheduling overhead and maximizing hardware utilization.
  • Engineered multi-agent orchestration pipelines, ensuring low-latency execution and scalability for enterprise-grade autonomous workflows.
  • Optimized compute-bound and memory-bound workloads by profiling execution on hardware accelerators, identifying/resolving PCIe and memory bottlenecks in distributed GPU infrastructure.
  • Designed end-to-end AI and data architectures for latency-sensitive edge environments, establishing architectural guardrails for business units and development teams to ensure scalable model deployment.
  • Optimized compute-bound and memory-bound workloads using CUDA profiling for scalable wireless analytics systems and distributed GPU infrastructure for AI RAN.
  • Engineered multi-agent systems orchestration for 5G/6G mobile core network architectures, utilizing Agent-to-Agent (A2A) communication patterns and the Model Context Protocol (MCP) to standardize AI model interactions across distributed edge nodes.
  • Integrated advanced AI/ML inference pipelines into enterprise C++ telecommunications processing systems, along with alignment-aware batching pipelines to improve Tensor Core memory access efficiency for GPU-accelerated inference workloads.

Patent Specialist - AI Systems & Architecture

• Nokia, Munich, Germany
• Duration: February 2026 - Present
• Description: Bridging the gap between abstract machine learning algorithms and physical silicon execution. Conducting deep technical evaluations of cutting-edge AI architectures, LLM inference pipelines, and agentic workflows to define and protect hardware-software co-design strategies.
• Core Responsibilities:
  • Architectural Deconstruction: Reverse-engineering and evaluating state-of-the-art AI inventions to determine structural patentability.
  • Technical Specification & Claim Engineering: Translating complex machine learning architectures into rigorous technical claim charts and engineering specifications for the European Patent Office (EPO).
  • Agentic System Evaluation: Leading the technical assessment of autonomous Agentic AI systems, defining the structural and ethical frameworks for AI-generated inventions.
  • IP Strategy for Compute: Leveraging a deep engineering background in bi-directional LLMs and ML classifiers to assess the technical viability and physical silicon impact of next-generation AI patents.

Student Research Assistant

• Technical University Munich, Department of Informatics, Chair of Network Architecture and Services
• Duration: February 2017 - August 2019
• Topic: Software support for the management of student projects (Chairman), Videocube

Master Thesis Student

• King's College London, London
• Duration: July 2018 - May 2019
• Topic: Content placement and pricing in connected edge networks. Engineered optimized data placement and routing algorithms for edge computing environments, directly addressing latency constraints in distributed telecommunication networks.

Student Research Assistant

• Forschungsstelle for Zahnrader und Getriebebau (FZG), Technical University of Munich
• Duration: October 2017 - October 2018
• Topic: Nextcloud system administration, osTicket system migration, PowerShell automation scripts

Research Intern

• Technical University of Munich, Munich, Germany
• Duration: March 2017 - July 2017
• Topic: Analysis of code-based cryptosystems (Supervisor: Prof. Antonia Wachter-Zeh)

Technical Consultant & System Developer

• Nectar Consultancy Services Pvt. Ltd., Kolkata, India
• Duration: July 2015 - June 2016
• Topic: Back-end developer of existing Mobile Text Based Internet Service

Summer Intern

• Indian Statistical Institute (ISI), Kolkata, India
• Duration: May 2014 - July 2014
• Topic: Facial gesture recognition with emotional parameters

Summer Trainee

• Bharat Sanchar Nigam Ltd. (BSNL)
• Duration: 2013
• Topic: Advanced telecommunication engineering

PhD (Dr.-Ing., Doktor der Ingenieurwissenschaften)

Technical University of Munich (TUM), Munich, Germany (Nov 2019 - Dec 2022)

• Department: Computer Science
• Supervisor: Prof. Dr.-Ing. Georg Carle
• Chair: Chair of Network Architecture and Services, Department of Informatics
• Dissertation: Coordination, Trust and Orchestration of Cognitive, Open, Multi-Vendor Network Automation Functions
• Doctoral research highlights:
  • Engineered real-life implementation for cognitive autonomous networks (multiple independent AI agents).
  • Developed decentralized AI control loops for automated telecommunication network management and operations, bridging theoretical ML with deployable 5G infrastructure.
• Additional coursework (TUM): Introduction to Docker; Parallel Programming Systems (PPS); Advanced Computer Architecture (ACA).

M.Sc. (Master of Science in Communications Engineering)

Technical University of Munich (TUM), Munich, Germany (Aug 2016 - May 2019)

• Supervisor: Prof. Dr.-Ing. Habil. Carmen Mas Machuca
• Department: Chair of Communication Networks, Department of Electrical and Computer Engineering
• Dissertation: Game theory based content placement and pricing in connected edge networks

B.E. (Bachelor of Engineering)

Jadavpur University (JU), Kolkata, India (2011 - 2015)

• Department: Electrical, Electronics and Communications Engineering
• Supervisor: Prof. Sudipta Chattopadhyay
• Dissertation: Generation Of New Sets Of Fixed Length And Variable Length Spreading Codes For CDMA/Optical CDMA System