Technical Consultant at IMT · Winnipeg, Canada

Dinkar
Taneja

Technical Consultant

Technical Consultant at IMT. Master data consulting, mainly for large health organizations.
IBM InfoSphere MDM set up, integrated and tuned, so every person has one trusted record.

✉️ Get in touch

Identity solutions delivered

A master index holds one trusted record for each person. All systems that hold that person use the same record.

Master Patient IndexHealthcare
Master Person IndexHuman Services
Master Citizen IndexeGovernment
Master Provider IndexHealthcare providers
5+

💼 Years on client projects at IMT. Started as Technical Analyst. Now Technical Consultant.

85%

🎯 Billable target. Passed again and again.

01

What I do

Ten areas of work. Each one shows what the work is, how it is done, and why it matters. Tap a row to read more.

Development mainly in Java and Spring Boot. Build custom solutions for what each client needs. The Spring Boot work fits with the client's own systems, shaped to their needs, so the client gets what they need.

JavaSpring BootCustom solutions

Master data consulting, mainly for large health organizations. Includes IBM InfoSphere MDM set up, shaped to client work. Configuration covers the data model, the services and the matching setup. When it runs well, every other system can trust the records it holds.

Data modelServicesMatching setup

Build inbound and outbound integrations between MDM and client systems. The links run in real time, in batches, or as services. Records flow in, MDM matches them, and clean records go back to the systems that use them. The integration work uses IBM App Connect Enterprise and IBM MQ. Some clients use Apache Kafka. Some clients still use legacy brokers.

Real timeBatchService basedIBM App Connect EnterpriseIBM MQApache KafkaLegacy brokers

One person may show up as J. Smith, John Smith or Jon Smith. The engine scores each pair. Matching rules, weights and thresholds are configured and tuned, then tested with threshold analysis. Good settings raise match accuracy. They also cut the review list for data stewards, so only the doubtful cases need a human decision.

Master data consulting for large health organizations, IBM MDM set up included. Identity solutions include Master Patient Index and Master Citizen Index. Each solution gives the client a single source of truth. Every system that holds that person uses the same record.

Use AI tools in day to day project work to raise productivity. AI helps write documentation, create scripts, support coding and build test cases. Routine work finishes faster, and more time goes to real client problems.

DocumentationScriptsCoding helpTest cases

Set up MDM on client servers and on cloud. Every client environment is different, so the setup follows what the client needs. PowerShell and Python scripts build environments the same way every time. No one has to build them by hand.

AWSMicrosoft AzureGoogle Cloud PlatformPowerShellPython

Help clients move from their own infrastructure to the cloud. Support upgrades and migrations. One example is a client application that moved from WildFly and Java 8 to Spring Boot on Java 25.

Provide implementation support and training to client teams, plus support after go live. Troubleshoot issues in matching, services and integrations until the system runs clean. Good support builds strong client relationships, and strong relationships bring repeat business.

Implementation supportTrainingPost go live supportTroubleshooting

Explain MDM solutions such as linking and matching to clients in plain words. A client should not need technical words to see the value. When the value is clear, decisions get easier.

02

Matching, in practice

This demo has three cases. One engine checks all three. The score shows what happens next. The system may act. A data steward may act. Or no one acts. Pick a case. Then press run.

Pick a case. Then run the engine

This is a common case. One person has three spellings of the name. The name changes a bit each time. The address and email also change a bit. But the strong proof is the same. The date of birth is the same. The postal code is the same. The ID is the same.

Registration system✓ linked

J. Smith

Date of birth
1988-03-14
Address
421 Elm St
Postal code
R3C 1A5
Email
j.smith@mail.com
Phone
Not on file
ID · last 3
••• 472
Billing system✓ linked

John A. Smith

Date of birth
1988-03-14
Address
421 Elm Street
Postal code
R3C 1A5
Email
Not on file
Phone
204 555 0182
ID · last 3
••• 472
Support portal✓ linked

Jon Smith

Date of birth
1988-03-14
Address
421 Elm St.
Postal code
R3C 1A5
Email
john.smith@mail.com
Phone
Not on file
ID · last 3
Not on file
Proof and points for the score
Name0
Date of birth0
Address0
Email & phone0
Postal + ID0
Match score0
Engine log
// Waiting to run. Press Run match
Where the score lands. The limits decide

🎚️ Try it. Set the limits yourself

11.0
13.0
J. Smith ↔ John A. Smith14.2✓ Linked by the system
A. Patel ↔ Amit Patel12.8→ Steward review
Margaret Chen ↔ Peggy Chen12.1→ Steward review
R. Dubois ↔ Robert Dubois10.4✕ Not linked
John Smith ↔ John Smith Jr.8.6✕ Not linked

With these limits, 1 pair links on its own. 2 pairs go to a steward. 2 pairs do not link. The case above has a score of 14.2. It links on its own.

The system links it on its own

The score is 14.2. It is above the limit of 13 for auto link. Three records become one linked view. No steward task is made. No person spends time on it.

Why the limits matter
Before
Stewards check too many cases

When the limits are not set well, too many cases go to stewards. Some cases are clear matches. Some cases are clear non matches. Stewards still have to check all of them. That takes many hours.

The work
Set the limits with care

The engine is tested with real pairs of records. The limits sit where the proof is strong. High scores link on their own. Middle scores go to a person to check. Low scores do not link. Try the sliders above. That is the job.

After
The system does the clear cases

The system now does most of the work. It links clear matches on its own. It keeps clear non matches apart. Stewards see only the hard cases. That takes less time. It costs less. Links happen fast.

This demo uses sample records. It uses a simple scale from 0 to 15. In real work, we set the limits for each client. We do this work in IBM MDM.

03

Selected work

Examples from client work at IMT. Client names are not shared. That is how client work should be. Each example shows the problem, the fix and the result.

🏥 MDM · Healthcare

One person, three records

The problem
The same person was in many systems many times. No one knew which record was full and correct.
The approach
Matching ran in IBM MDM. Close pairs went to stewards to check. The system did not join them on its own.
The outcome
Now there is one linked view for each person. Stewards keep watch for new copies.
🏛️ Integration · eGovernment

A new system joins with no mess

The problem
A new system had to send records to MDM. It used its own formats. It also used its own ID numbers.
The approach
Fields from the new system were mapped to MDM. Records flowed in, got matched, and went back clean. Counts were checked at each step.
The outcome
The new system went live. Old matches still worked. Problems showed up in test, not in live use.
👥 Matching · Human Services

Why do these records not match?

The problem
The system missed real matches. It linked wrong records. Stewards had too much work to check by hand.
The approach
Found the fields that confused the engine. Changed the rules. Set new limits.
The outcome
The steward list got much smaller. Now stewards see only the hard cases.
Personal project

DataMapper

Data mapping tool built with Spring Boot in Java 25. A small AI helper suggests field matches and cuts the manual effort. Data stored in SQL Server.

Spring BootJava 25AI helperSQL Server
04

Experience

Close to six years with IMT in Canada, in data consulting for large health organizations and other large groups.

Jun 2024 to Present

Technical Consultant

IMT. InfoMagnetics Technologies Corporation. Winnipeg, Canada

Master data consulting, mainly for large health organizations. IBM MDM set up and integration.

  • Build the integrations between MDM and client systems.
  • Tune matching rules, weights and thresholds.
  • Deliver identity solutions. Master Patient Index, Master Citizen Index, Master Person Index and Master Provider Index.
  • Set up MDM on client servers and on AWS, Azure and Google Cloud Platform.
  • Help clients move to the cloud. Support upgrades and migrations.
  • Use AI tools in day to day work for documentation, scripts, coding and test cases.
  • Support and train client teams, before and after go live.
  • 🎯 Pass the 85% billable target again and again.
Jan 2021 to Jun 2024

Technical Analyst

IMT · Winnipeg, Canada

Worked on MDM each day. Helped set up and support client systems. Used SQL in this work. Strong base for the consultant role.

↑ Grew into Technical Consultant in Jun 2024
Apr to Jul 2017

Intern

Adwitiya · India

First real project work, done while studying for a degree. Real deadlines. Other people used the code.

2015 to 2017

Social Media Specialist & Coordinator

India

Start of the work story. Grew Facebook pages to 600k+ likes. The work taught a key skill: explain hard things in a clear way, in simple words.

05

The toolbox

Tools and skills, grouped by the job they do. Tap a tag and the note on how it is used opens right under its group.

🎯 Master data. The daily work

🗄️ Data & databases

🔌 Integration

💻 Development

☁️ Cloud & platforms

Pick a tag

Tap any tag above to see how it is used in the work.

06

About

At IMT in Winnipeg since Jan 2021, close to six years on client projects. Started as a Technical Analyst. Technical Consultant since Jun 2024. Master data consulting, mainly for large health organizations, with IBM MDM set up as part of the work. Human services and government work come next.

Degree in Computer Science from DAV College in Jalandhar. AEC in IT Programming Analyst from LaSalle College, Montréal, finished in 2020. Now living and working in Winnipeg.

The work helps the data of a group agree with itself. Each person gets one trusted view. All systems that hold a record for that person use that same view.

Jalandhar, India→ Montréal, Canada→ Winnipeg, Canada
Always ongoing

Studying and improving skills

Self study to keep learning, with a strong focus on AI and new technology. Knowledge built beyond the day to day job.

2020 · MONTRÉAL

AEC in IT Programming Analyst

LaSalle College. Programming, databases and systems. A strong base for all the work since then.

INDIA

Bachelor's in Computer Science

DAV College in Jalandhar. The start of the work with computers. First internship at Adwitiya during that time.

07

Outside work

When I am not at work, this is me.

Badminton

I play whenever I can. Fast feet, fast smashes, and if there is a court I am on it.

🏆 District Champion, Under 17State level representation, Under 19

Cooking

I love to cook healthy food at home in Winnipeg. I try a new recipe on most weekends. Weekend chef at home.

Outdoors

Hiking and outdoor challenges with friends, making the most of the short Winnipeg summers.

Blood donation

I am a certified donor.

Always a learner

Study on my own to keep learning, with a strong focus on AI and new technology, building knowledge beyond the day to day job.

Believes in AI

Keeps learning AI, and more and more work gets done using AI.

08

Take a break

There are two short games. Each game has its own full screen. The page stays clean. You get a big space to play. Pick one game. Play as long as you like.

Game 1

Steward’s Call

You play as the data steward. You see two records. You also see one match score. You make the call. You can link the records. You can send them for a check. Or you can keep them apart. Some cases are traps. That is just like real work.

60 seconds. Streaks score big. Best score: none yet

Game 2

Match the Records

This memory game comes from my daily work. The same person is written in two ways. Flip the bright cards. Match each pair of records for the same person. One card may use a short name. One may use just a first letter. One may have a small spelling mistake. Each matched pair gets its own colour.

8 pairs. 8 colours. Win with the fewest moves.

✉️ Get in touch

Technical Consultant at IMT in Winnipeg. This site shows the day to day work. Questions about data or MDM, or just a wish to connect, are welcome. Fill in the form and press Send. The message goes straight to the inbox.

Based in📍 Winnipeg, Manitoba, Canada
Find meLinkedIn · GitHub. Search for Dinkar Taneja
LanguagesEnglish, Hindi, Punjabi
Winnipeg timeLoading time in Winnipeg

Prefer your own email app? Open this message in your email app instead.

At the end of the day, it all has to agree. Drag a card. Then link them into one.

Single trusted record
James Smith
Born 14 Mar 1988 · Winnipeg, MB
Merged from 4 sources. Match score 14.2. Linked by the system.
4 records hanging. All are the same person.