Pradeep Dwarakanath

CTO

Bengaluru, Karnataka, India25 yrs 9 mos experience

Most Likely To SwitchHighly Stable

Key Highlights

Led ML and DNN applications for Qualcomm's AI program
Developed DSP algorithms for satellite communication systems
Authored multiple patents and conference papers

Stackforce AI infers this person is a Telecommunications and Wireless Communication expert with a strong focus on signal processing and machine learning.

Contact

Skills

Core Skills

Machine LearningDeep LearningDigital Signal ProcessingSatellite CommunicationTelecommunicationsData ProcessingData OptimizationWireless CommunicationReceiver DesignWireless NetworkingSignal ProcessingSystem AnalysisStandardizationEmbedded Systems

Other Skills

ML frameworksDNN frameworksTensorflowPytorchCaffe2ONNXDSP AlgorithmsControl softwareHSPAData rate improvementsDual SIM Dual Standby3G Dual SIMData throughputBottleneck analysisSuper Regenerative Receivers

About

1) Lead of ML and DNN applications for Qualcomm's AI inference accelerator program 2) DSP Firmware lead for Qualcomm's Satellite communication modem project 3) 3G Data plane and throuughput end to end (AP & CP) handling 4) Design & Architecture of 3G Protocol stack with bottleneck analysis and optimization 5) Call and Throughput KPI improvement 6) Receiver architecture for Ultra low power receiver (802.15.4Q based devices) 7) PHY Algorithms for Multi-User MIMO based Wireless LAN (802.11 N & AC) 8) System analysis of Coordinated multi-point MIMO system for LTE Release 11 9) System analysis of MIMO system for LTE Release 10 10) WCDMA Power control & measurements 11) GSM/GPRS TDMA processing engine 12) 10 Quality patents that have multiple citations 13) 2 Internal and 1 external conference paper submission

Experience

25 yrs 9 mos

Total Experience

6 yrs 5 mos

Average Tenure

8 yrs 10 mos

Current Experience

Qualcomm

2 roles

Head of Machine learning and Deep neural Networks applications

Promoted

Oct 2018 – Present · 7 yrs 7 mos

Working on ML / DNN frameworks for AI inference accelerator program
Development of import layer for frameworks like Tensorflow, Pytorch, Caffe2, ONNX, etc..
Development of Operators and kernels for DNN for various frameworks
Development of Neural Network models and validation

ML frameworksDNN frameworksTensorflowPytorchCaffe2ONNX+2

DSP firmware lead for Satellite Modem project

Jul 2017 – Oct 2018 · 1 yr 3 mos

Development of DSP Algorithms for Satellite communication system
Developed highly optimized algos like DL Equalizer, Frequency tracking, time tracking and other control SW

DSP AlgorithmsSatellite communicationControl softwareDigital Signal ProcessingSatellite Communication

Samsung, bangalore

3G and HSPA DSP lead

Sep 2016 – May 2017 · 8 mos · Bangalore

Data rate improvements on HSPA during Dual Sim Dual Standby Operation
Reduction of tune-in/tune-away times
Regular feature updates and issue handling

HSPAData rate improvementsDual SIM Dual StandbyTelecommunicationsData Processing

Samsung

5 roles

UMTS Data Path redesign and Dual SIM Dual Standby development

Jan 2013 – Sep 2016 · 3 yrs 8 mos

3G Dual Sim feature with end to end ownership on Data throughput
Improved the data throughput from a prototype level to a commercial quality phone
In-depth bottle neck analysis of DSDS 3G throughput led to 5-6 cross layer enhancements
These Cross layer optimizations improved data rates by 12/ 20% for average/ best case
Redesign of UMTS Data path was carried out to gain run-time efficiency, maintainability and scalability
Optimized RLC processing and gained 5% run time efficiency. Lead to data rate improvement by 8 Mbps in Dual carrier scenario
1 Grade “A” Patent
1 accepted paper in IEEE CCNC 2016

3G Dual SIMData throughputBottleneck analysisTelecommunicationsData Optimization

Ultra low Power communication for Body Area Network (BAN) devices

Jan 2012 – Dec 2012 · 11 mos

Conducted study of Super Regenerative Receivers (SRR) and their suitability to BAN (802.15.4Q)
Derived performance equations for in-Band, Adjacent & Alternate Band Interference & Noise
Devised a new Quench waveform that improved the SRR’s Frequency selectivity there-by improving the Adjacent and Alternate Band Interference rejection
Assisted Prototype development with theoretical framework and simulation model to analyze SRR’s sensitivity to Interference and noise and suitability to (802.15.4Q based devices)
1 Grade “A” patent in improving Adjacent & Alternate carrier frequency response for SRR

Super Regenerative ReceiversFrequency selectivitySimulation modelsWireless CommunicationReceiver Design

Baseband algorithm development for 802.11n & 11ac

Jan 2010 – Dec 2011 · 1 yr 11 mos

PHY Simulation bed development for 802.11n & AC 4x4 MIMO system
SU and MU MIMO processing, Receiver processing (Channel Estimation & Equalization)
Time & frequency offset estimation and correction, RF and PHY Impairments related processing
Compressed and quantized feedback
Precoders like Dominant Eigen mode, Block diagonalization, SLNR minimization precoder etc.
Validation using results from peer contributions from TGn & TGac members & IEEE papers
2 Grade “A” Patents

802.11n802.11acMIMO processingWireless NetworkingSignal Processing

Analysis of LTE Release 8 and CoMP System

Jan 2007 – Dec 2009 · 2 yrs 11 mos

Development of CoMP simulation framework in “C” to assist standardization
Literature survey for selection of scheduling algorithm, CoMP Scheme
Implementation of SFN based precoding CoMP and studied system performance
Interference modeling, Channel estimation
MIMO equalization with interference suppression Multi cell PMI & CQI feedback
PF Scheduling, Precoding, HARQ, Transmit chain
Channel estimation, PMI & CQI feedback on pilots, equalization

CoMP simulationScheduling algorithmsMIMO equalizationTelecommunicationsSystem Analysis

WCDMA/HSPA L1 controller commercialization & Standardization

Jan 2003 – Dec 2006 · 3 yrs 11 mos

Commercialization of Samsung’s HSPA phones with contribution in 3G L1 controller
Inner & Outer loop power control; Power optimization in RACH module
L1 part of cell selection & reselection criterion and framework, Soft/ softer handover
Measurements related to cell selection and handovers
Closely followed HSPA work items like Enhanced FACH and Continuous packet connectivity
Produced 3 standards contribution
3 Grade “A” & 2 Grade “B” patents

HSPA commercializationPower controlCell selectionTelecommunicationsStandardization