Sayandev Mukherjee

Sayandev Mukherjee Email and Phone Number

Co-Founder and Head of AI @ MeshSage
California, United States
Sayandev Mukherjee's Location
Santa Clara, California, United States, United States
Sayandev Mukherjee's Contact Details

Sayandev Mukherjee work email

Sayandev Mukherjee personal email

About Sayandev Mukherjee

2019-: Applications and use cases of AI/ML to communication networks in general, both wired and wireless, from the physical layer all the way up to applications and even standards design.2018-2019: Machine and Deep Learning applied to wireless networks -- MLaaS platform for a wireless network: model training, deployment, and lifecycle management, from data ingestion to feature extraction to model training to deployment to performance monitoring.2017-2018: Enhancing the efficiency of deep neural network training algorithms and improving the robustness of deep learning image classifiers to adversarial attack. I use various deep learning frameworks, including Keras, TensorFlow, and PyTorch.2013-17: Data scientist and researcher applying machine learning algorithms and structured knowledge bases to natural language text/documents. I'm mostly working on integrating algorithmic modules into processing pipelines for novel services. I used a mixture of shell script, Java, and Python to write these pipelines.[Before 2013, exclusively; since 2013, at about 10%]: Wireless technology researcher with over 15 years of experience in research and development of cellular and other wireless technologies, including TDMA, CDMA, and OFDMA access technologies. Systems I have worked with include IS-136, GSM, IS-95, cdma2000, W-CDMA, 3G1xEV-DO, HSPA, LTE, and IEEE 802.16.Specialties: My most recent work in wireless has been in applying results in stochastic geometry (some of which were original to me) in the analytical modeling of multi-tier heterogeneous cellular networks. I have written two monographs on this subject (one solely authored by me, the other co-authored by me). I have worked on resource allocation problems in wireless systems, where by "resource" I mean time slot scheduling, code allocation, power allocations, etc. Most recently, I have designed, implemented, and lab-tested algorithms for auto-configuring networks of access points under the coordination of a central controller, covering all aspects of operation from installation onward. My work has involved a mix of analysis and simulation. I have written my own simulators in Matlab and C/C++.

Sayandev Mukherjee's Current Company Details
MeshSage

Meshsage

View
Co-Founder and Head of AI
California, United States
Sayandev Mukherjee Work Experience Details
  • Meshsage
    Co-Founder And Head Of Ai
    Meshsage
    California, United States
  • Cablelabs
    Distinguished Technologist
    Cablelabs Nov 2022 - Present
    Louisville, Co, Us
    Applying Machine Learning to communication networks, wired and wireless.
  • Cablelabs
    Principal Architect
    Cablelabs Sep 2019 - Nov 2022
    Louisville, Co, Us
    Pursuing exciting applications of AI and ML to wireless communication systems
  • Huawei Technologies
    Principal Machine Learning Architect
    Huawei Technologies Jul 2018 - Sep 2019
    Shenzhen, Guangdong, Cn
    Machine Learning as a Service (MLaaS) platform for 4G and 5G Wireless Networks● Designing support for data augmentation, feature extraction, model training, deployment, and monitoring under the topological, compute, and storage constraints of a wireless network● Federated Learning, Split Learning, and hierarchical distributed decentralized training methods● Neural Architecture Search methods
  • Docomo Innovations, Inc.
    Senior Research Engineer
    Docomo Innovations, Inc. Jul 2013 - Jul 2018
    Palo Alto, Ca, Us
    Deep Learning -- enhancements to training algorithmsDeployed and evaluated several hyperparameter optimization algorithms for training deep neural networks (including Hyperband and Bayesian optimization), focusing on image classification modelsProposed, deployed, and evaluated methods to make image classification deep neural network models robust to adversarial examplesMachine Learning for natural language-initiated task recognition and implementationImplemented, trained, and tested machine learning methods to extract task information from natural language text for applications to service offerings on mobile devicesData Ninja Services (https://www.dataninja.net): a commercial suite of text analytics APIs (closed in October 2017)Designed, implemented, and deployed several APIs in the Data Ninja suite of servicesDesigned overall data workflow to create indexes and dictionaries used by all services, including automated periodic updatesDesigned and implemented prototype news exploration web service with machine learning based automatic topic extraction and semantically meaningful topic labeling
  • Docomo Innovations, Inc.
    Senior Research Engineer
    Docomo Innovations, Inc. Apr 2010 - Jun 2013
    Palo Alto, Ca, Us
    Wireless communications - research and standardsHARQ in the LTE and LTE-A standard: wrote a Matlab simulation platform; simulated the performance of the LTE HARQ specification; proposed and simulated the performance of other HARQ retransmission schemes with different definitions of the redundancy versionsDownlink SINR analysis in heterogeneous cellular networks, with multiple tiers of base stations, such as macro/micro/pico/femto (both open and closed access): derived results on outage and camping probabilities for arbitrary users in such heterogeneous networks
  • Spidercloud Wireless, Inc
    Senior Staff Design Engineer
    Spidercloud Wireless, Inc May 2008 - Mar 2010
    Milpitas, Ca, Us
    A startup founded to make an enterprise-RAN comprising a network of 3G femto-access-points with a central controller, providing 3G data rates with high user densities within the enterprise, all without requiring any modification (software or hardware) to users' handsets.Radio Resource ManagementDesigned, implemented, and lab-tested algorithms to auto-configure a network of access points under the coordination of a central controller, including: (a) measurement of radio environment by access points to allow for automated choice of spreading codes; (b) automated assignment of transmit powers to access points based on standards-mandated measurements from user equipment; (c) periodic updating/refreshing/aging out of old measurements; and (d) handling sudden changes to the network such as the insertion or deletion of access pointsDesigned simulation platform to compare the performance of a network with central coordination versus a collection of consumer femto-cells without coordination, for different densities of cell placement and different sizes of deployment areas.Proposed a scheme to mitigate the impact on the uplink at a closed subscriber group cell of transmissions by a local user not authorized to access this cell (patent applied for)HSUPA Scheduler design and developmentDesigned and implemented a comprehensive simulation platform to compare the performance of mac-es schedulers on the uplink of a HSUPA network, with varying degrees of central coordination; proposed and simulated different scheduler schemes to reduce or eliminate spikes in received power on the uplink in the presence or absence of users in soft handoff, thereby enhancing system stability (patents applied for)Tested scheduler performance in a full-stack hardware HSUPA laboratory setup involving a controller, an access point, and several user terminals, including automated scripts to trigger soft handoffs, extract relevant data from log files, and plot them
  • Marvell Semiconductor, Inc
    Senior Staff Design Engineer
    Marvell Semiconductor, Inc Oct 2006 - May 2008
    Designed a simulation suite (in Matlab and C++) for a WiMAX system to compare the bit and block error rate sensitivities of various MIMO receiver signal processing algorithmsContributed the sensitivity results to the WiMAX standard's Radio Conformance Testing (RCT) requirements; also represented Marvell at IEEE 802.16 standards conferences
  • Bell Laboratories
    Member Of Technical Staff
    Bell Laboratories Nov 1996 - Sep 2006
    Murray Hill, Nj, Us
    Support for mobile users in the IEEE 802.16-2004 standard Proposed, analyzed, and patented protocols to (i) support user mobility (including handoffs) in the IEEE 802.16-2004 standard, originally intended for stationary users only, and (ii) reduce energy consumption and increase battery life of untethered users in such a system, all without making any change to the existing 802.16-2004 standard.An All IP network architecture for a Wireless Communications SystemDesigned the fast-cell-site selection and power control schemes for a comprehensive proposal for a distributed all-IP cellular data network with latency small enough to support voice over IP (VoIP). Advanced Wireless Access (AWA) System This was a fixed wireless loop-type system based on steerable antennas and Time Domain Multiple Access (TDMA) designed to provide high speed (several Mbps) data services in addition to voice.Designed and implemented a system simulator (in Matlab and C)Proposed dynamic frequency-allocation schemes to maximize frequency reuse and minimize interference in the system (patent awarded)
  • Bell Laboratories
    Other Projects
    Bell Laboratories Nov 1996 - Sep 2006
    Murray Hill, Nj, Us
    As a researcher in the area of wireless at Bell Laboratories, I have worked on resource allocation problems for TDMA (both IS-136 and GSM) and CDMA (IS-95, cdma2000, W-CDMA), and on capacity and coverage issues for relay, mesh, and ad-hoc networks. The work usually involved a mixture of analysis and simulation, with the simulators written in either Matlab or C/C++.Connectivity and Power Usage in Ad-Hoc and Hybrid Ad-Hoc Networks Ad Hoc networks--performance and topology modification algorithms Scheduling Strategies for Cellular Systems with Relays Analysis of Proportional-Fair scheduling algorithm and its application to opportunistic beamforming Comparison of average total downlink transmit power with and without soft handoff in UMTS Relationship Between Bit and Block Error Rates for Viterbi Decoders Role of Measurement and Prediction in Effective Handoff in Cellular Systems Handoffs in Cellular Systems Macro- and Micro-diversity in Cellular Systems Statistical Properties of Wireless Communication Channels Viterbi Equalizer for Digital PCS wireless handset

Sayandev Mukherjee Skills

Lte Wireless Wimax Algorithms Umts Mimo Simulations Cdma Wcdma Hspa Cellular Communications Matlab Ofdm 3g Wifi Gsm Mobile Communications 3gpp 4g Ofdma Network Architecture Wireless Technologies Femtocell

Sayandev Mukherjee Education Details

  • Cornell University
    Cornell University
    Electrical Engineering
  • Indian Institute Of Technology, Kanpur
    Indian Institute Of Technology, Kanpur
    Electrical Engineering
  • Kendriya Vidyalaya
    Kendriya Vidyalaya
    Aisse; Aissce

Frequently Asked Questions about Sayandev Mukherjee

What company does Sayandev Mukherjee work for?

Sayandev Mukherjee works for Meshsage

What is Sayandev Mukherjee's role at the current company?

Sayandev Mukherjee's current role is Co-Founder and Head of AI.

What is Sayandev Mukherjee's email address?

Sayandev Mukherjee's email address is sa****@****ail.com

What schools did Sayandev Mukherjee attend?

Sayandev Mukherjee attended Cornell University, Indian Institute Of Technology, Kanpur, Kendriya Vidyalaya.

What skills is Sayandev Mukherjee known for?

Sayandev Mukherjee has skills like Lte, Wireless, Wimax, Algorithms, Umts, Mimo, Simulations, Cdma, Wcdma, Hspa, Cellular Communications, Matlab.

Free Chrome Extension

Find emails, phones & company data instantly

Find verified emails from LinkedIn profiles
Get direct phone numbers & mobile contacts
Access company data & employee information
Works directly on LinkedIn - no copy/paste needed
Get Chrome Extension - Free

Aero Online

Your AI prospecting assistant

Download 750 million emails and 100 million phone numbers

Access emails and phone numbers of over 750 million business users. Instantly download verified profiles using 20+ filters, including location, job title, company, function, and industry.