coreviewsystems

[coreviewsystems]

5 ways to Improve Team Productivity

“I had told you to send minutes as soon as the meetings were over, why did you not do it?”

“You say that Sam had asked you to do it that way, but how can you prove it?”

“What was the testing team doing? How did they ok this build?”

The questions were flying fast, the meeting room had gone silent, the faces grim. This is a common scene when the latest build is not working as it was supposed to. The customer makes the escalation call up the escalation chain. All the seniors right from the top boss demand for the status every hour on the hour. The team lead suddenly has to handle 4 status meetings to attend, build that is failing, and no clue about what actually has gone wrong. We all have faced this situation. Anxiety levels rising, the blame starts passing from delivery team to QA team to Deployment team back and forth. Teams are intent on defending themselves, that they did not make a mistake — they followed the process.

How do you resolve this situation? How do you make them productive? Here is what has worked for me.

Communicate with customers and let them know the real situation and how and by when the problem will be solved setting realistic expectations. Keep them updated.

Focus the team on solving the problem at hand. Work with them to form a plan to detect and resolve the problem at hand.

Absorb all the pressure from the seniors. Take responsibility for the situation. Do not just pass it through to the team. A team under undue pressure will make more mistakes

Bring the status meetings to reasonable numbers. Let the doers do and not spend time on the report.

Have a root cause analysis meeting after the escalation is handled.

Let me know what worked for you in the comments.

To read more articles visit our website: www.coreviewsystems.com

[coreviewsystems]

Is Your Business Stuck?

“Makarand, we have been doing the same thing for the last few years. The business is stagnant. There is a total loss of motivation in the leadership team. I think we should drop what we are doing and do something else, that is completely different.”

How often did you face similar situations in your business or your job? Changing a job may be easy, but that may not be the best thing for your career. Changing a business is even more difficult. The feeling of getting stuck can add to the frustration.

Here are some tips for the entrepreneurs facing similar challenges.

• Evaluate your business from a financial perspective. Why would you want to move away from something that has already started paying off?

• Talk to your customers. If they are very happy, that is a good enough reason to continue. If they are not, you will know what needs to be fixed.

• Set reasonable expectations. Having unachievable goals can lead to frustration. Having lower goals can lead to boredom.

• Talk to those who have faced it before. You will get some invaluable tips.

• Look at restructuring your business, there are hidden opportunities that were neglected due to busy schedules.

• Take a strategic break.

Don’t think you are alone. Frustration is a very common feeling. The last couple of years were challenging for many.

How did you get out of it, the last time?

To read more articles visit our website: www.coreviewsystems.com

[coreviewsystems]

How to define your AI goals

Failure to define AI goals is one of the main reasons why companies fail to implement AI successfully.AI is only as good as the goals that are set for it. If you want to reap the benefits of AI, you need to first define your AI goals.

If you’re not sure where to start, this article will provide some tips on how to define your AI goals:

Develop problem-solving ability: Artificial intelligence systems can solve complex tasks by assigning them to reliable AI systems, simplifying critical tasks.

Incorporate knowledge representation: Knowledge representation reveals real-world information that a computer can use to solve complex real-life problems, such as diagnosing a medical condition or interacting with humans naturally. To achieve the desired results, AI models can be fine-tuned and optimized using the represented information

Allow continuous learning: AI solutions are fundamentally based on learning. Observations and past experiences enable computer algorithms to understand AI programs conceptually. In AI programs, input-output pairs are processed for a defined function, and the results are used to predict outcomes for new inputs.

Encourage social Intelligence: AI that recognizes, interprets, and simulates human feelings, emotions, and experiences is called affective computing, or ’emotion AI.’ The aim of affective computing is to allow AI systems to interact and socialize like humans by reading facial expressions, body language, and voice tones.

Promote creativity: With the help of artificial intelligence, humans can accomplish tasks more efficiently and creatively. In the future, artificial intelligence will be able to whittle down vast amounts of data, consider options and alternatives, and create creative paths and opportunities.

Promote synergy between humans and AI: In AI, one of the key goals is to develop a synergy between humans and AI so they can work together and enhance each other’s capabilities rather than depending on one another.

To read more articles visit our website:www.coreviewsystems.com

[coreviewsystems]

5 good work habits that will set you apart

It is November 21st. The yearend is approaching fast. It’s the time when everyone is publishing lists. Trends for the future, top ten things to watch out and so on. I am going to take this opportunity and make a list of top five habits that have helped me over the years.

Build a framework of concepts in your head: The principles of computing are always applicable, so whatever language / framework / technology you use — they will always deal with the principles. There would be differences in syntax or implementation details. If you learn to correlate the principles from one language to another — it becomes easier to learn new ones.

Do not think about implementation details when you are designing. Design to fulfil the requirements. The implementation details should come later. There could be limitations to implementation but the design should not have limitations.

Learn to read code — Reading code is the most important skill one could have. It helps in all manners. It teaches you to how to write and how not to write code. It helps you debug the problem. It makes you a better reviewer — the list can go on.

Appreciate and strive to write elegant code — I think that as any craftsman takes pride in their handiwork — we as programmers take pride in the elegance of our code.

Comment and Document — You will forget why you took certain decision, or how to configure given system, so always document it where it can be searched quickly. Build a system to remember.

What is your list of habits?

To read more articles visit our website: www.coreviewsystems.com

[coreviewsystems]

Organization growth-Things to consider

“My organisation is very nimble, and I have to keep changing things regularly in response to the customer requirements.” Amit was right in his argument. Any organization growth will change based on many external factors, customers are the most important ones.

When the senior leadership keeps changing the path, the people at lower levels need to understand the reasons and the impact of the change.

The time it takes to make the change effective will depend on the size of the organization. The leader’s ability to think about the long term impact plays a very important role in scaling the organisation. The one who depends on instant changes and insists on seeing an immediate change will not be effective in running a larger organisation.

Here are some things to consider when the organizations grows rapidly

• Have a clearly defined vision and mission. Write it down and make it known to every employee. Ensure that the changes you are making are in line with the vision. Frequent deviation from the defined vision can lead to chaotic growth.

• Work on preparing the next line in command. Involve them in evaluation and implementation of the change

• Get external help, it helps to work with newer vendors who can support you in the growth.

• Get advice from people who have done it before. Every situation is unique, so make your own evaluation of the advice given, before following.

At a certain point the leader clearly sees the inability to manage the growth. The proven effective way to manage this situation is to transfer the responsibility to someone else, who may be better suited.

Are you facing growth challenges?

To read more articles visit our website: www.coreviewsystems.com

[coreviewsystems]

Making the most out of a forced holiday

I have come across two types of #holidays

The first one I call vacation. #vacation is planned time away from work. This is time set aside for enjoyment with friends and family. During vacation I prefer to keep completely away from work, being available only for real emergencies, and I have never come across any real emergency while on vacation. Have you?

The second type is a set of forced holidays due to festivals or similar calendar events. I have absolutely no control on these holidays. As far as possible, I avoid travelling during these days. There are many who love to stand in long queues or get stuck in traffic jams, but I am not of that kind.

These holidays offer a perfect opportunity for introspection. I call them strategic breaks. In the course of routine work activities, there is hardly any time left for a long term perspective. Keeping away from the routine work, I take a look at my work as an outsider. Looking at it from a distance shows things that are not visible on a day to day basis.

During one such break earlier, me and my partners decided to focus on building expertise in machine learning. Over the few years, we implemented multiple solutions for our customers using machine learning that became a key driving force during our recent acquisition by Suma Soft.

Any suggestions on what I should do next week?

To read more articles visit our website: www.coreviewsystems.com

[coreviewsystems]

Related Data Loading using EF Core — Part 1

In few of previous posts we have seen how Transaction management can be handled using EF Core. Let’s now understand different ways by which data can be loaded using EF Core.

There are basically 3 common patterns to load the related data. 1. Eager Loading 2. Explicit Loading 3. Lazy Loading

This is Part 1 of related Data Loading patterns and we will cover Eager Loading in this part.

Eager Loading in EF Core :

Eager Loading is a pattern of loading related data along with the main entity. This means that the main entity and related data can be fetched using a single query saving the time and bandwidth.

Eager Loading is done using the “.Include()” method as shown in following example :

using (DBEntities context = new DBEntities()) { var employees = context.Employees.Include(x => x.EmpDetails). ToList();

foreach (var emp in employees) { string EmpName = emp.Name; string EmpCode = emp.Code; string EmpAddress = emp.EmpDetails.Address; } }

Eager Loading using “.AutoInclude()” method is introduced in EF Core 6.0 : Auto Include helps include the related data every time the entity gets loaded. It has same effect as calling “.Include()” in every query to get the results as shown in following example :

modelBuilder.Entity<Employees>().Navigation(e => e.EmpDetails).AutoInclude();

To read more articles visit our website: www.coreviewsystems.com

[coreviewsystems]

Token-based authentication

In the modern digital era, the way web applications are designed and developed has come a long way compared to traditional web applications. Accessing applications or resources is no more limited to computers. Rather mobile devices are being used widely to access websites and resources. The resources are not necessarily private in nature — and modern-day applications even extend certain resources to be accessible by external entities.

And in all such scenarios, security is of utmost importance. The resources being shared with the requestors must ensure authentication and authorization are implemented correctly protecting resources from being attacked by malicious users.

Token-based authentication provides a reliable and robust solution to ensure secure access to the resources within the network or beyond the private network boundaries. Based on the level of security, suitable options available — hardware tokens, software tokens, one-time passwords, biometric identities — can be selected. Usage of multi-factor authentication combined with tokens can create strong secure applications which are not easy to target by hackers/malicious users. Tokens not only offer better security but also enhance the user experience to transition thru multiple resources seamlessly — making the life of end users easier.

Token standards like OAuth/JWT are being widely adopted by developers due to ease of implementation along with a defined structure of the same in web applications/services. Tokens have inherent advantages like being stateless, having defined validity, and encrypted — authentication and authorization can be integrated within your application landscape easily. Technically — tokens like OAuth/JWT are typically stored on the client side without storing sensitive information on the same. This benefits the load balanced/fault tolerant environments to use resources like servers/applications with the same tokens — which otherwise would have meant centralized session management for servers in most of the traditional web applications.

With multiple customers, we have utilized OAuth tokens extensively for REST APIs. JWT was also used for cross-application integrations. If you are thinking of enhancing the security of your application landscape by consolidating the implementation of your authentication and authorization needs — utilizing tokens can be a good strategy to consider.

To read more articles visit our website: www.coreviewsystems.com

[coreviewsystems]

Organizations: how to deliver a customer-centric approach

The current business climate is changing very fast due the global exposure and technological advances. The organizations that are embracing technological advances are thriving in these times. On the other hand, customers have become savvier and are expecting a lot.

How can the organizations deliver the customer-centric approach?

They can do so, by anticipating what the customers want, and when they want it, with the help of data analytics.

As the first step to taking help predicting customer needs and behavior, all the customer-related data is cleaned and accumulated on a modern data platform. The importance of the quality of data cannot be emphasized enough. I have mentioned it in my previous posts as well. If the data quality is not good, it can negatively affect the analysis.

Once the data is gathered, it can be effectively used to know the customers, their behavior, needs, and preferences. Data mining and predictive analysis can be used to find the hidden trends, identifying the high potential customers and the offers they could be interested in this increases customer acquisition at a reduced cost.

It can also help in retaining existing customers and reactivating older prospects.

To read more articles visit our website: www.coreviewsystems.com

[coreviewsystems]

Human Learning and AI

It was my first month in the new job. I was working with a new prospect and visited them multiple times to finalize the proposal for their requirements. It may sound a bit strange now, but in those days we used to meet in person for discussions. Phones were used only to get appointments. It was becoming a routine for me to arrive at the company guesthouse the previous night and to leave for the office early in the morning just after a quick breakfast.

This time I arrived very late at night and did not get a chance to inform the caretaker about my next day’s plans. I got ready early in the morning and reached the kitchen hoping to get something for breakfast, I was up for a big surprise. My favorite breakfast was already waiting for me. As I was close to finishing the food a hot cup of tea with just the right amount of sugar was delivered to my table.

Sounds like a great story of customer service? Yes sure it is, and it is also a great example of human learning.

The dependence on human learning fails miserably when we implement it at a large scale and in different geographies. There must be some way to automate it using machines.

Why not? Look at this –

Recently, when I was getting late to start the office, my phone assistant reminded me and also offered to find a different route to make up for the delay. I was surprised. I had not set any reminders. A lot of research is already in place, making it feasible for regular use, and I will not be surprised if I see more such examples soon.

Have you tried to use AI/ML for business benefit?

To read more articles visit our website: www.coreviewsystems.com

[coreviewsystems]

Machine Learning with Graphs

Complex domains have a rich relational structure which can be represented as a relational graph consisting of entities and relationships. Graphs are a language built on graph data structures to describe and analyze entities with relations/interactions.

Many types of data are graphs.

Social networks

Communication networks

Internet

Knowledge Graphs

Research paper citation networks

Molecules

And many more.

We can take advantage of the relational structure by explicitly modeling relationships amongst entities through graphs as conventional deep learning approaches are designed for simple sequences and grids. Networks are complex with arbitrary size and complex topological structure, no fixed node ordering, dynamic, and have multimodal features.

Traditional ML for graphs pipeline uses hand-designed features. Given an input graph, extract node, edge, and graph-level features, and learn a model like Random forest, SVM, Neural networks, etc. that maps features to labels.

Graph representational learning alleviates the need to do feature engineering every time. After representing the data in graphs with entities, relationships, and their respective properties/features, nodes(entities) are mapped to d-dimensional embeddings such that similar nodes in the network are embedded close to each other.

Different types of Graphs ML tasks:

Node level

Node classification

Examples: Categorize online users/items

Edge level

Link prediction

Examples: Recommend items users might like

Graph level

Graph classification

Examples: Drug discovery

2. Clustering

Examples: Social circle detection

These Graph machine learning tasks lead to high-impact applications in different domains.

We at CoreView see data from different perspectives and have created graph representations of data for multiple problems wherever applicable to solve complex data use cases in various domains.

Have you ever come across any such use cases where you could have improved the outcome by leveraging Graphs?

To read more articles visit our website: www.coreviewsystems.com

[coreviewsystems]

What is Intelligence?

If intelligence is your ability to make decisions based on available information, then it should be the easiest task for computers. There is no lack of information, or data as we say it, and no lack of processing power either. Then what makes it so difficult for the computer to develop intelligence?

The basic problem is that the term intelligence itself is not well defined. When I looked up on the web to find out what it means, I was flooded with so many diverse definitions, that I decided to sum it up with my own understanding.

Being a goal-oriented person, the definition I could relate to the most was the one formalized by Shane Legg & Marcus Hutter — “Intelligence measures an agent’s ability to achieve goals in a wide range of environments.” The key here is that the environment will change in an expected as well as unexpected way as we start working.

– Gathering information

– Filtering information based on what could be useful

– Tagging and storing information for future use

– Restructuring based on what we are looking for

– Computing inferences and assigning weightage to them

– Evaluate what-if scenarios based on different inferences and actions

– Execute actions and measure effects

– Provide feedback to the system (retrain) to increase the effectiveness

The ones that are difficult to quantify include

– The ability to take risks

– The ability to evaluate the impact on human life

One extraordinary ability that human intelligence brings in is the ability to learn from other people’s mistakes.

Remember any example of learning from others?

To read more articles visit our website: www.coreviewsystems.com

[coreviewsystems]

Creating A Robust Infrastructure To Utilize Data Streams

It is raining heavily in my city. It reminds me of the data that is produced or recorded by even a moderately sized organization. Numerous data streams are flooding the business every day. In order to harness this force, the business has to judiciously create the data lakes and dam the data without harming the business environment, using the data gathered to gain insights, predictions, and directions from it. Ok, let me stop this analogy before we head towards a cloud burst! J

A smart organization builds a robust infrastructure to utilize this data.

The first step is to collect the data from where it is produced, it could be the recording of client calls, data gathered from instrumentation, sensors, logs, or external systems. The instrumentation and infrastructure engineers would be working at this stage

The second step is to transform and store this data in a usable format, be it structured or unstructured. Start cleaning and Prepping the data. This is done by the data engineering team

The third step is to aggregate and label the data. Experimentation, A/B testing, Training data, and using ML algorithms. This task is done by data analysts and data scientists.

Each one of these steps requires a different type of training and expertise. The boundaries between the steps are porous and the people working on them have overlapping skills. So it is imperative that the correct person is assigned the correct role. Not realizing this can result in unhappy scientists and underutilization of the actual skills.

It is best to match the steps and skills correctly.

To read more articles visit our website: www.coreviewsystems.com

[coreviewsystems]

Machines learning with a focus

On the first day in the office at my first job, I reached much before others. Banu, manning the reception, was the only person around. She asked me to wait in the reception area.

After a short while, other employees started arriving. I was observing how they were entering the office, how they were dressed, how they were noticing and greeted each other, the tone of their conversation, and their use of lockers. I can list many. Just by observing, I was learning many things about the office environment.

“Hi Makarand, how are you doing? My name is Harsh, and I am your buddy in the office.” I could feel the warm welcome. Harsh became my friend in no time. “Come on in, I will show you the office.” Harsh introduced me to many people and places. I became much more comfortable. “Here is a phone number to reach me. Feel free to call anytime. Would you like to join me for lunch today?” Harsh left me in a large conference room, where I completed joining formalities.

On Monday next week, all new employees were summoned to a conference room for induction training. The training covered many other aspects of office working and employee behavior which empowered us to become responsible and efficient employees.

Knowingly or unknowingly I went through the process of • Self-learning when I was on my own • Assisted learning, when I was with the buddy • And Focused learning in the classroom

Human learning is a gradual process. Machine learning is much faster. Most machine learning exercises use focused learning for specific objectives. For example, when we wanted to come up with specific policy recommendations for prospective customers, we focused on learning buying patterns from earlier customers along with various factors that impact their buying decisions.

Are your machines learning with focus?

To read more articles visit our website: www.coreviewsystems.com

[coreviewsystems]

Using transactions in entity framework core

A database transaction is a sequence of multiple operations performed on a database, and all served as a single logical unit of work. If the transaction is committed, all of the operations are successfully committed to the database. If the transaction is rolled back, none of the operations are committed to the database.

In Entity Framework, the SaveChanges() method is responsible for internally creating a transaction that wraps all INSERT, UPDATE, and DELETE operations under it. Multiple SaveChanges() calls, create separate transactions, perform CRUD operations and then commit each transaction.

Default Transaction Behavior :

By default, if the database provider supports transactions, all changes in a single call to SaveChanges are committed in a transaction. If any of the changes fail, then the transaction is rolled back and none of the changes are committed to the database. This means that SaveChanges is guaranteed to either completely succeed, or leave the database unmodified if an error occurs.

Multiple SaveChanges in a Single Transaction :

Entity Framework Core and Entity Framework 6 provide support to create or use a single transaction with multiple SaveChanges() calls using the methods mentioned below:

a) DbContext.Database.BeginTransaction(): This method creates a new transaction for the database and allows us to commit or roll back changes made to the database using multiple SaveChanges method calls.

b) DbContext.Database.UseTransaction(): This method provides support to pass an existing transaction object created out of the scope of a context object. This will allow EF to execute commands within an external transaction object. Alternatively, pass in null to clear the framework’s knowledge of that transaction.

Savepoints :

When SaveChanges operations are performed multiple times within a transaction, Entity Framework core automatically creates a savepoint for every SaveChanges call. In case SaveChanges() fails due to any reason, the transaction can be rolled back to the last successful Savepoint. After coming to the last savepoint, the transaction is in the same state as if the remaining part of the transaction has not been executed yet. Savepoints can be created manually and have user-defined names.

Usage of Savepoints :

a) CreateSavepoint(“NameString”): This method creates a savepoint in the transaction. This allows restoring the transaction state to what it was at the time of the savepoint creation.

b) RollbackToSavepoint(“NameString”): This method rolls back all commands that were executed after the specified savepoint was established.

Before Dot NET 3.5, Developers often used to write ADO.NET code to save or retrieve data from the database. Developers had to create Database connections and maintain them for various CRUD operations according to the business requirement. This was a cumbersome and error-prone process. Microsoft introduced a framework called “Entity Framework” to automate all these database-related activities for your application and its first version was released around 2008. Entity Framework Core 6 also known as EF Core 6 is the latest version of Microsoft’s Entity Framework (as of the publishing date of this post). It is an Object Relational Mapping (ORM) framework that is an enhanced version of ADO.NET. This ORM automatically handles the data storage and retrieval from the database. EF Core is very powerful and also very easy to learn and use in your projects.

To read more articles visit our website: www.coreviewsystems.com

[coreviewsystems]

The dilemma: Cloud computing or Edge computing?

Sensors, controllers, gateways, MQTT, AMQP… all these buzzwords welcome you to the ‘connected’ world of Internet-of-Things and its inseparable buddy, Cloud-computing! Poor sensors, at the heart of the system, have little idea of the intricate cobweb built around the measurements diligently taken by them, synonymous with the heartbeats in the living beings!

The sensors emit measurements, there are nearby ‘smart’ devices to capture and resend those to the bigger system lying in the cloud. The cloud collects the data and various services turn the data into meaningful information. This typical IoT workflow came with its own side-effects — data explosion and compute-hungry applications!

However, IoT is evolving into more mature design patterns… ‘Edge-computing’ is such a pattern that tries to perform some computation on or near the device. The advantages? Bandwidth, storage space, and computing power are saved in the cloud. It’s essentially shifting some of the ‘processing intelligence’ to the device.

Well, so far so good! Is edge-computing eliminating the role of cloud computing? No way! There are scenarios when edge computing is beneficial, and other scenarios when cloud computing is the way to go…

Typically devices send similar measurements with occasional spikes or anomalies in them. It’s a waste of resources to process similar messages ~99% of the time. In edge computing, the messages are summarized/aggregated on the device itself, or on a nearby controller, while they are in the normal range. The summary data is sent to the cloud at a less frequency but if the data shows some anomaly, it is sent immediately to the cloud. This especially helps save resources if the data is in binary form — audio, video, or images.

Real-time, on-the-field actions or notifications are another problem that edge-computing solves elegantly. Suppose you have a system for monitoring cameras that should beep an alert if there is any intrusion. This is best handled on the on-premises controller, rather than flowing the image to the cloud which then sends out a notification.

Data privacy concerns also find a solution in edge computing. Hospitals that are apprehensive about the privacy of patients’ data can benefit from a design based on edge computing.

On the other hand, Cloud computing has an upper hand when data processing needs a lot of ‘elastic’ aka auto-scaling computing power, also as a permanent home for the data. There is no alternative to cloud computing for the cold-path-data-analysis — analyze the trends for the last year, capacity planning, and forecast of the growth.

All-in-all cloud computing is a broad umbrella giving shelter to the intelligent and mechanical devices in the world of IoT… Do you agree with me?

To read more articles visit our website: www.coreviewsystems.com

[coreviewsystems]

Be no 1 in whatever you do

“We are just starting to develop; how can you expect us to be at number one from day one? There are other players in the market. They are working for a few years and way ahead of us.”

It was a natural reaction from Yash to my demand. Every startup must start afresh, which obviously means following the competition. Right?

Being no 1 is a mindset and not a position. Creating that mindset within the team can be done from day one. There are ways to reach that position sooner than you expect by following these simple methods.

Redefine that finish line. Number one does not always mean the revenues or the market share. Can you identify some other criteria? This will let you get to the position sooner. Once you get there, redefine the criteria, so that getting to the number one position again becomes an archivable goal.

Define a restricted market. Can we reach number one in a smaller market segment? The segment expands automatically as you get there.

Identify niche areas that are underserved. Can we become the best provider in any of that areas?

These are just examples.

Can you think of any other simpler way?

To read more articles visit our website: www.coreviewsystems.com