Big Data, Technology & Integration, QA

Treselle Systems help clients develop products and execute IT projects in the areas of big data, cloud and enterprise applications. We also have deep expertise in database management (SQL & NoSQL), technology integration, UI development, and testing services. Treselle Systems is an award winning technology services and product development company head quartered in US with offices in Campbell, CA, and development center in Chennai, India.

 
$5,000+
 
$25 - $49 / hr
 
50 - 249
 Founded
2004
Show all +
Chennai, India
headquarters
  • Chennai 600024
    India

Portfolio

Key clients: 

Discern

Adopt a Pet

American Tours International

American Media Inc

Quality Assurance of Retail Stores Data in Motion

Perform quality checks on retail store data between source data in MongoDB and transformed Master Data in MySQL.

Perform quality checks between Master Data in MySQL and JSON cache files for accuracy.

Perform random checks on large retail stores plotted on Google maps and compare against the API payload response.

Capture screenshot of Google maps with markers at random to check the accuracy of retail store marker plots.

Generate dashboard with different reports such as total stores, open/close stores, history of stores over the past few months, and others between source data in MongoDB and master data in MySQL for internal curators to perform sanity checks.

Reduce QA manual process time and automate checks across different systems.

Provide confidence to the business users that data persisted across systems such as MongoDB, MySQL, and Cache file systems, and the data plotted on Google maps are in sync.

...

Anomaly Detection

Identify anomalies & outliers beyond traditional heuristic algorithms and static business rules in a stream of events coming from weblogs, partner integration as JSON payload, and data vendor integration in XML and other formats.

Design & Runtime Ecosystem:

Anomaly Detection Techniques & Methods

Anomaly Detection Process Proposal

Design & architecture with different technologies & framework (on-premise & on-cloud)

PoC

https://www.linkedin.com/pulse/anomaly-detection-poc-4-member-team-1-spr...

Regression & Multi-class prediction for Capital Market

Ingested 15 minute delay stock ticker feeds via Kinesis Firehose into S3.

Lambda was configured on the S3 data events and notifies EMR job that performs cleansing, filtering, data wrangling, calculate derived attributes from base attributes, and derive necessary features needed for the Machine Learning.

AWS ML is continuously trained on the ML model that uses both regression and multi-class algorithms for prediction purpose.

Multiple endpoints are created on these ML models and necessary DNS is associated with Route53 to perform predictions on the data submitted by the end users.

Internal curators and business analyst use BI tools via Athena connector.

AWS Cognito is used to manage user pools to enable only subscribed users can perform real time predictions from the web platform.

https://www.linkedin.com/pulse/our-amazing-...

Social Facial Detection App

Designed & architect a micro service API backend entirely on AWS using 95% managed services for fail-over and scalability.

Implemented facial recognition using AWS Rekognition managed service with custom scoring algorithm to decide whether the interested parties should be communicated or not.

Implemented other AWS services such as Lambda, DynamoDB with DAX, Logging with Cloudwatch, ElasticSearch & Kibana, Cloudfront, CloudFormation, S3, Glacier, SES, SNS, and API Gateway.

https://www.linkedin.com/pulse/our-amazing-journey-amazon-web-services-r...

Capital Market Analytics

Grew from 2 to 40 + engineers that includes technical architect, project managers, backend engineers, data engineers, data science engineers, database developers, frontend engineers, and QA engineers. 

Same team worked on multiple sectors such as Healthcare, Energy, Retail and Real Estate that required deep context switching across these multiple sectors to develop and release within the same 2-week sprint cycle.

Implemented complex statistical distributed computation, geo-spatial analysis, text mining, natural language processing, and machine learning models for peer comparison, backtesting, forecasting, entity matching, force & rank, trend analysis, predictions, rollups to multiple US geographical regions, etc.

Released close to 4 MVPs across multiple sectors for customer demos, roadshows, and investor meetings to get feedback from the customers and investors.

Designed one of its kind Energy data model with master & transaction data by...

Data Matching - Entity Identification, Resolution & Linkage

Automate massive scraping of retail stores and reduce the overall processing time from 45 days to 1 week
Automate and Optimize store matching, identification, and linkage to tag the status of each store by checking with historical data that was done using excel
Automate 80 data quality checks that were done manually by 45 data quality analysts which was one of the main bottlenecks in the pipeline process
Improve the accuracy of geo-location of retail stores as downstream analytics rely heavily on these attributes
Automate & Improve the accuracy of stores mapped to appropriate banners for proper disposition analytics
Reduce overall headcount of Mechanical Turk Force by 85%

https://www.linkedin.com/pulse/mechanical-turk-force-optimization-automa...

Reviews

Sort by

Long Term Dev Partnership for Analytics Firm

"They have earned our business over and over again."

Quality: 
4.5
Schedule: 
5.0
Cost: 
5.0
Willing to refer: 
5.0
The Project
 
$1,000,000 - $9,999,999
 
Nov. 2012 - Ongoing
Project summary: 

Treselle designed and developed a search tool and interactive analytics capabilities, as well as a monitoring system for a data marketplace company.

The Reviewer
 
11-50 Employees
 
San Francisco Bay Area
Harry Blount
CEO, Discern
 
Verified
The Review
Feedback summary: 

Their efficiency and readiness enable the acceptance of large projects on short notice. Additionally, Treselle is proactive, with ideas and suggestions that have saved thousands of dollars.

 

BACKGROUND

Introduce your business and what you do there.

I am the CEO of Discern, which is a data marketplace that connects business decision makers to the data they need to inform their decisions.

OPPORTUNITY / CHALLENGE

What challenge were you trying to address with Treselle Systems?

We needed a partner that could both take the edge off the volatility of our growth and provide highly skilled resources, which are hard to come by here in Silicon Valley.

SOLUTION

What was the scope of their involvement?

Treselle helped build and design an intuitive search engine and embedded interactive analytics capabilities, as well as a swarm of intelligent agents to notify users of significant changes in the data. They worked with structured and unstructured databases and also utilized the Amazon Cloud. Their feedback and discussion surrounding particular tradeoffs have also been valuable.

What is the team dynamic?

There is one primary project manager and a secondary one who steps in when needed.

How did you come to work with Treselle Systems?

When our Vice President of Engineering joined the company about five years ago, he suggested the structure of a small core team in the US and a larger, more flexible partner offshore. After about a three-week vetting process he narrowed the considerations to three candidates, flew to see the Treselle team in Chennai, and concluded that had all the necessary components.

How much have you invested with them?

We are currently investing about $45,000 per month with Treselle, but that number has fluctuated over time.

What is the status of this engagement?

The engagement began in November 2012 and is ongoing.

RESULTS & FEEDBACK

What evidence can you share that demonstrates the impact of the engagement?

Somewhat frequently we experience unexpected surges in business from large customers with quick deadlines and multiple requirements, and Treselle has always been prepared for those instances, which is what enables us to say "yes." They always come through to help meet our deadlines regardless of personal circumstances, and therefore we feel that they have earned our business over and over again.

How did Treselle Systems perform from a project management standpoint?

The project management is very good. Every three months we rotate sending some of our people to Chennai and them sending people here because we think that face-to-face time is very valuable. In the intermediate term, there are usually sprints planned out two or three in advance, for which we provide requirements and they write understanding documents as a check to ensure we are on the same page. On the short term, we have daily phone calls and status updates. For project management tools we use both Jira and Zendesk.

What did you find most impressive about them?

Treselle is very proactive in making suggestions to make the workflow more efficient and effective, which has probably saved multiple months worth of time and money that would have been used re-engineering. That characteristic is what makes them feel more like a partner and less just a vendor.

Are there any areas they could improve?

Not that I can think of, no.

Do you have any advice for potential customers?

I would encourage any company going into business with an overseas partner to invest the time to meet face to face. Once we did that and locked in the processes, the partnership really began to flourish.

5.0
Overall Score We feel like we knocked it out of the park when we found them as a partner.
  • 5.0 Scheduling
    ON TIME / DEADLINES
    They deliver on time. If they say they're going to do it, they do it.
  • 5.0 Cost
    Value / within estimates
    From a cost perspective they are competitive.
  • 4.5 Quality
    Service & deliverables
    Only because we aren't always perfectly on the same page.
  • 5.0 NPS
    Willing to refer
    There's no doubt in my mind that I would refer them.

Outsourced Data Management & Development Team

"If I were to start another company, Treselle is the standard I’d compare everyone to."

Quality: 
5.0
Schedule: 
5.0
Cost: 
5.0
Willing to refer: 
5.0
The Project
 
$200,000 to $999,999
 
Nov. 2012 - Ongoing
Project summary: 

Treselle tackles processing for diverse data sets after having assumed development of existing software, website, and AWS infrastructure. They scrape and repackage data using machine-learning techniques.

The Reviewer
 
11-50 Employees
 
Chandler, Arizona
Troy Hansen
Co-founder, Discern
 
Verified
The Review
Feedback summary: 

They deftly handle every challenge with skill sets rivaling big data firms. While turnover requires occasional adjustment, they expertly modify team size, allocate resources, and offer direct access to each member. Their competent leadership and constant communication build trust and peace of mind. 

BACKGROUND

Introduce your business and what you do there.

We’re a financial data company. We purchase data from different sources, process it, and package it to customers’ requirements. The most important part of our business is how we transform it and mix-and-match it. We create daily or more frequent reports, and our customers can access data via our website in any way they want. Our customers are Wall Street companies that use our data to drive their bottom line.

I am a co-founder of the company, as well as the head of QA, the release manager, and the data manager.

OPPORTUNITY / CHALLENGE

What challenge were you trying to address with Treselle Systems?

We don’t have U.S. programmers, and we needed help with our data processing solution. We had been working with a Chinese firm and had an existing platform in place.

SOLUTION

What was the scope of their involvement?

Treselle worked on our data-processing software and website, taking over our Amazon Web Services infrastructure. They handle data scraping and processing, site security, and every other related aspect.

Our company is organized into sectors, with [SMEs] subject-matter experts in each who use proprietary methods to decide what data we will ingest and how we will manipulate it. Based on their input, I write requirements and send them to Treselle’s team, who implement and deploy them on a test platform. Then I perform QA. If the results meet our SMEs’ requirements, the work is promoted to our real, customer-facing platform.

Treselle has had to deal with different types of data and technologies. In the energy sector, we process every oil well in the U.S. It’s all public-domain data being scraped from state websites with tools their team developed and implemented. For real estate, our data comes in various formats—Oracle spreadsheets, CSV or other flat files and MySQL databases.

Treselle also wrote an elaborate NLP [natural language processing] system for drug and device trials in the health care sector. Our SMEs designed meaningful keyword searches, and we applied algorithms to determine how meaningful a text was. That enabled us to extract drug trial data and show secondary variations of the trials to customers.

Once we have the data, we organize and clean it through algorithms they’ve implemented. They use various machine-learning techniques, like linear regression models, predictive models, and clustering tables. All of this had been done in Excel spreadsheets, but it’s now automated using code. As customers send in parameters, we apply algorithms and produce results.

What is the team dynamic?

They’ve had up to 30 engineers working for us and as low as 10, depending on our development status.

We have a lead technical person, Raghavan [Madabusi, VP Engineering]. He has two people directly under him. They manage the project and day-to-day operations. They may be designated as backend and frontend, but I can go to either one and get what I need.

Under them, the team is broken down into different areas, including database, UI/UX, gurus, and in-house QA. There is a lead for each group, with resources under them.

How did you come to work with Treselle Systems?

I don’t know.

How much have you invested with them?

I believe we’re spending around $50,000 per month on their services, but I don’t know if this has been a consistent number.

What is the status of this engagement?

We started working with Treselle in November 2012. We had consulted with them before that time.

RESULTS & FEEDBACK

What evidence can you share that demonstrates the impact of the engagement?

Treselle has done some cutting-edge programming over the years, using machine-learning technology typically seen in big data.

As an idea of what we process, our Discern data model for all the data in the system has over 2 billion points in it at the moment. Our daily ingestion is over 1 million points of data.

How did Treselle Systems perform from a project management standpoint?

As a startup, we’ve appreciated their flexibility in changing the team size. They seem to have a lot of developer turnover, so it can take a few days to find the right person, and at times we’ve had to adjust priorities accordingly. This is only an issue for one-off tasks; they’re perfect in terms of everyday routine. In terms of breadth of work, we’ve given Treselle a huge variety of things to do, and they’ve always been able to find someone who could get it done.

I like that I can talk directly to the tech leads, the management, and the actual developers. I am an engineer at heart, so I want to go to the source and get it done. Treselle is OK with this, unlike other companies I’ve used. I’m connected with all of their employees on Skype, and we use JIRA to track problems and requirements. The team submits comments and questions for me via the platform.

We have 2-week development cycles, with a points system assigned to each task. They assign the points at the beginning, based on the requirements we write. We prioritize tasks accordingly, and they start a sprint. We may disagree about the numbers, but not very often. After all these years, we’ve come to understand what it takes. We receive daily updates on their progress for each sprint.

I interface with one of my two main contacts every night. When we get into problems, resolutions, bugs, or requirements, they set up a conference with us and the technical lead for the group we’re working with, in order to eliminate phone-chain problems.

What did you find most impressive about them?

Recently, we needed an entirely new demo on an emergency timeline. They made that happen, despite the fact that there was a cyclone in their part of the world and public transportation was down.

They handle communication well. At no point in our sprints have I been left wondering what they were working on or if they had a problem. That contributes to peace of mind, especially since they’re on the other side of the world. When I have to report on a feature or customer to our CEO, I can give him exact data. If it’s not what we want, I can tell the team and alter priorities to get what we need.

Another strength is that I can give them generic requirements and trust them to tease out the important details. They help me get it right the first time and get to the root of what we want, largely because of their leadership. Both of my contacts have worked in the States before and they know what it takes to build a product.

Are there any areas they could improve?

I have a good working relationship with them, but I do remember one low point. I had spent a lot of time mentoring one of their new QA team members on our process. She became very thorough and one of the best QA people I’ve ever worked with. After 6 months, she relocated and had to leave the company. There was nothing Treselle could do, but it was very unfortunate to lose that time I invested.

5.0
Overall Score They’re much better than our previous Chinese team.
  • 5.0 Scheduling
    ON TIME / DEADLINES
    They’re available 18 hours a day, even when not at work.
  • 5.0 Cost
    Value / within estimates
    Treselle is less expensive than U.S. teams and just as good.
  • 5.0 Quality
    Service & deliverables
    They’ve gone above and beyond.
  • 5.0 NPS
    Willing to refer
    If I were to start another company, Treselle is the standard I’d compare everyone to. They’re very good.