Building the Future of Monitoring with Artificial Intelligence

Incedo IncBuilding the Future of Monitoring with Artificial

Intelligence

Confidential and Proprietary - Incedo Inc

Our Second Acquisition– increased focus on specialized skills

Launched Incubation Lab for clients

• Specialized Product Engineering Services and Data & Analytics Competency

• Expertise in Financial Services, Life Science and Communication Engineering

• Focused on Emerging Technologies and Innovation - in the context of our clients’ businesses

• Bay Area Headquartered - 1500 employees across North America, South Africa and India (Gurgaon, Bangalore)

• End-to-end capabilities across Application Services, Infrastructure and Operations

• Agile, responsive engagement models

2011 2012 2013 2014 2015 20160

200400600800

1000120014001600

Empl

oyee

s

Incedo Spun out of a diversified $4B Conglomerate

Started US & South Africa Operations

Started Life Sciences and Data & Analytics practice

• 35+ clients• Growth of 75%+ y-o-y

Our First Acquisition - grew Product Engineering Services

Started work with US based Tier 1bank

We Help Businesses Excel And Solve Complex Problems With Technology

Incedo Overview

2

Success is proven by growth of more than 671% since its inception in 2011

Incedo in the list of Top 10 Emerging Analytics Start-ups for the year 2016

Our Offices•North America•Santa Clara, New Jersey & New York

•South Africa• Cape Town & Johannesburg•Asia-Pacific (India)• Delhi, Bangalore, Chennai & Mumbai


We are building a special technology firm and the world has begun to take noteIncedo is an INC 5000 awardee – fastest growing private companies in North America

Incedo recognized by Dataquest for 85% growth

Incedo is one of Top 10 Emerging Analytics Start-ups for the year 2016

Incedo is one of 20 Most Promising Tech Service Providersby CIOReview

Incedo ranks #16 on CRN’s Fast Growth 150

3Confidential and Proprietary - Incedo Inc

Feasibility Go-No go

Pilot Project/ Joint Investment

Full Scale Development

What We Do How We Do it

Frameworks

Proof of Concept

Accelerators

Follow Technologies

Track Trends

Rapid Prototyping

3D Visualization Framework on Mobile devices Machine Learning for Edge Devices Master Data Management Solution Accelerators

Accelerators/ Frameworks

IoT Incubation Labs at IncedoWe Partner With Our Clients to Build Smarter Solutions for Them

4 Confidential and Proprietary - Incedo Inc

AI & NLP Solution – Voice Based BoT Framework

Incubation Labs IoT & Connected Devices AI, NLP, Machine Learning Mobility NFV & Cloud Virtualization

IOT EVOLUTION AND FUTURE TRENDS Incedo’s Perspective of IoT Market

Confidential and Proprietary - Incedo Inc5

Internet has come a long way since 1980s but now it has entered into the most exciting phase

6


Driven by proliferation of connected devices & sensors

Monitoring and Control

Multi-transport and Multi-function Devices, Sensors

High Speed Wireless – 4G, 5G

Delivering Experience to Consumers

Connectivity and Mobility

Converged, Secured and Collaborative Operations

Digitized

Connected

Services Enabled

7


8

Number of people vs number of connected “things”

By the end of this decade 212 billion “things” are going to be connected to the internet

2003

0.12010

1.82020

30

Consumer60%

Industrial40%

2015

3.5Connected Devices Per Person

We are here, 99.9% of things are still unconnected


2020 View : Human Centric IOT - “4th Wave, 1.5T+Users/Things”Enabling Man-Machine Collaboration led by Robotics & Artificial Intelligence

Open Innovative eco system

SensorsDevicesMachines

System of

Engagement

System of

Records

Benefit To

Humans

Human Empowerment

Machine Intelligence Connected

Infrastructure

9


10

IoT has many dimensions… and is a Journey… Emerging Industrial IoTIoT Foundation

Disconnected Connect & ControlConnect

Product Smart Product Smart Connected

Product

Connect | Operate | Optimize

Remote Monitoring & Diagnostics

Extend to Ecosystem

Smart Factory – Asset Tracking | Predictive Analytics |

Robotics |Artificial Intelligence

Industrial operations are increasingly connected and monitored with connected devices. Exabytes of data is being generated by these devices, but only a fraction of it is being used in a meaningful way to derive insights and take real-time actions. What is next in industrial evolution?


Future of Industrial IoT would be Led by AI & Machine LearningiRobot Romba

Sentrian - Disease deterioration modeling platform

Amazon Echo

Artificial Intelligence is turbocharging industries that are generating massive volumes of data through their IoT investments.


CASE EXAMPLE : COMPUTATIONAL INTELLIGENCE FOR INDUSTRIAL IOT

Incedo’s Experience : Future of Monitoring Led by Artificial Intelligence


13

Continuous Monitoring

Safety

Operational Visibility &

KPIs

Predictive Maintenance

Performance &

Optimization

Asset Utilization

Real-time ResponseFast Data : Sub-second

Long Term AnalysisBig Data : Months-Years

Biggest IIoT challenge: deliver business value?

Customers: Early stages of IoT


14

Industrial safety solution40+ TB per year

120GB /day of sensor data

Industrial safety solutionBig Data: 40+ TB per year Fast

data: sub-second alerts

Hybrid engine manufacturerBig Data: 720 TB data

Fast Data: 20M Events/Sec

Benefits:• Operational Safety & efficiency • Regulatory compliance.

Benefits• New revenue & services • Predictive maintenance

14

IIoT business value is driven by both big & fast data

Safety

Performance &

Optimization

Managing sensor data is key to the value driven by Industrial IOT. Data from 50 billion sensors will generate 35 ZB annual data. 60% of this data is going to be generated by Industrial IOT.


100’s sensors

Edgegateway

15

1000’s sensors

Edgegateway

HighCost

Massive Data

LowBandwidth

Cloud Analytics

FastData

BigData

Ingestion Rate100,000+ data points/second

1000’s of locations

IIoT pain with big and fast data

Programming updatesContinuous learning

Amount of data generated by industrial location is very large and coupled with low BW connectivity presents unique challenges. 20 billion terabytes of industrial sensor data will be generated annually by 2020 * Intel/Cisco


2,000 Engines 5,512 Sensors

each2 Samples/sec

Incedo’s Solution – Machine Learning & Data Compression

720 TB/Year $1.8M on 4G LTE

Edge Analytics

Not enough Compute

Programmingtakes months

Real time lossless data

40x Compression

Deploy in days

Solution builder tradeoffs - Drop data or face increased cost or forgo- real time analytics

Machine Learning for Edge (Partner Solution)

Sensor Health &Cleansing

Data Accelerator(Data

Compression)Anomaly Detection


Solution is 2 fold - 1st is crunch the data at edge so that your can enable latency sensitive applications that require real time or near real-time response and 2nd is to enable sending of more data to cloud so you customers can have a global view on analytics and leverage their existing big data investments .

CASE EXAMPLE : NLP FOR SMART OPERATIONS



Text, Voice or Personae – Incedo’s Guided Autonomous IntelligenceOverview of Incedo’s BOT Solution Framework


BOT – Solution which can respond “intelligently” to input (voice and/or text)Intelligence = Input parsing + Context + Knowledge database+ Interactive Output

• Voice/Video to text• Language parsing

o N-gram analysiso Entity Identification

• Building Knowledge Databaseo Supervised Trainingo Unsupervised Training (Machine Learning)

• Context Managemento Domain Specific knowledge

• User Experienceo Responsiveo Richo Interactive

Input to entities

Entities to format ContextDictionary

NLP Engine

Text

Web

Voice

SMS

App

RSS

Management Interface

Domain Unsupervised dataTypical Voice/Chat BoT Application

User Inputs

• Supervised Configuration• Analysis

Incedo has experience in all elements of BoT solutions for Industrial, Healthcare and Communication Engineering

19

Personal Chat Assistance for Smart OperationsCase Example : Leading US-Bank required a Robotic Personal Chat Assistant that can be integrated with different internal applications within Banks Framework. The Interactive chat BOT eased the use of reporting platform by addressing common queries and requests of internal employees using existing reporting frameworks in the backend.

Developed solution over Python NLTK with Web, IoS and Mail based interfaces


Front End

Core EngineReport

Generator Simulator

Look Up

Data

Trained ModelNLTK Model

1

3 4 5

2

Sentence tokenization,pos tagging

Report Dictionary

6

Classification

Rest API

Raw Text

User InputEg: Filter & Output type

Augmented workforce for efficient engineering operations

ASR Engine

Service BOTs in Smart Operations which can answer simple speech based questions for routine operations, identify items, search inventories, act as a guide and even summon experts for a video chat

Please confirm the order “Pick up at Bay1, Inventory Lot 5…and Drop point: Bay 2, Loading No.7….


An example of speech based routine operations for shipment of materials

CASE EXAMPLE : DIAGNOSTICS AS A SERVICE



“Smart Car” IoT Service – PoC Overview• Telemetry Data Agent (TDA) uploads car data to the

IoT backend service

• IoT Backend service (AWS IoT) detects abnormal car health conditions (Accidents, Low Battery, Engine & coolant temp, Fuel )

• IoT Backend service notifies Car users and 3rd party service providers about abnormal conditions

• Portal (UI) provides Traffic view, Hazard declaring view, Car service Vehicle analytics view and Hospital Accident/Emergency service views

• Mobile App is used for Registration purpose, View his/her Car health stats, get notified on bad health conditions, get notified if family/friend vehicle has met with accident and details/status of hospital attending. Raise alarm by the owner if need emergency hospital or service need attention

Telemetry Data Agent (TDA)

Car Data

3rd Party Service provider RESTful Applications

Notifications

Mobile App

IoT Backend Service (AWS IoT)

Diagnostics As a Service for Aftermarket Cars Maintenance


Incedo Automated Testing Framework (ATF) for Industrial IoTOverview


Incedo Automated Testing Framework (ATF)

SUT Adaptation Engine: Tools/solutions to adapt ATF to SUT solution space

Framework Core (Configuration, Execution, Log Analysis , Reporting Engine; Core Libraries)

Test Scripting Engine (Scripts, Data, Suite)

ReportingEngine

Database

Highlights- Automated testing framework

(Single Click)- Single Node as well as End-2-End

Solution Testing capabilities- Flexibility in adapting to solution

domain- Wide range of reporting,

charting and trend capabilities- Ease of integrating of

opensource/ commercial test solutions into ATF

- XML based scripting- Low/No ramp-up time required

for scripting

Network under Test

Test Tool A Test Tool B

Test Tool CSUT

SUT

SUT

SUT

Case where SUT is a multi-node solution; Test Tools can be tools for injection and/or probe

www.incedoinc.com

[email protected]

Thank You.

24

NoteThe content and data shared in the proposal is Incedo proprietary and is intended for the party with whom the presentation is shared. It cannot be shared, copied or redistributed without Incedo’s formal consent.


http://www.incedoinc.com/

mailto:[email protected]