Ajay Bidyarthy’s Case for Starting Small

Ajay Bidyarthy has spent years working on large problems. His advice about data infrastructure begins with a modest instruction: “start small.” In a 2018 article, he tells organizations with less than five terabytes of data to concentrate on making their information queryable and choosing a business intelligence tool. The elaborate machinery can wait. For a founder selling analytics and software expertise, it is an appealingly restrained opening. Even the person who can build the complicated thing allows that you may have no use for it yet.

That practical question - how much machinery does a problem actually need? - offers a way into Bidyarthy’s career. Before founding Blackcoffer in 2016, he studied mathematics and computing, researched network security, and worked on algorithms for finding approximate solutions to difficult optimization problems. He later helped build a networking platform organized around physical proximity. Across those different settings, his work repeatedly concerns making something large enough to be confusing usable enough to act on.

A childhood between Jamui and Kolkata

Bidyarthy identifies Jamui, Bihar, as his native place and says much of his schooling and early development happened there and in Kolkata. Those are the beginnings he chooses to put in his public professional biography, alongside a fairly extensive menu of mathematical interests. The geography matters without requiring a heroic origin story. His route to running a Delhi-based consulting business passed through more than one city and more than one kind of technical work.

He attended the Indian Institute of Technology Guwahati from 2009 to 2013, studying Mathematics and Computing. His full name, Ajay Shankar Bidyarthy, appears among the students who fulfilled the requirements for the B.Tech degree in that discipline. The combination is a useful introduction to what followed: an education in both describing a problem formally and getting a computer to do something about it.

His student interests included simulation, optimization, operations research, and quantitative risk analysis. These are fields concerned with choices under constraints. There may be limited time, limited capacity, competing objectives, or more possible arrangements than anyone could sensibly inspect. Such problems have an unfashionable habit of remaining difficult regardless of how enthusiastically the meeting has gone.

Before the dashboards, the algorithms

One early research subject was data exfiltration: the unauthorized movement of information out of a network. In 2011, Bidyarthy co-authored Behavior model for detecting data exfiltration in network environment with Rajamenakshi Ramachandran and Subramanian Neelakantan. His early career also included network security research at C-DAC. Here, the task was to recognize troubling behavior inside the flow of ordinary activity, a demanding exercise in separating a meaningful signal from background movement.

Operations research at TCS’s Tata Research Development and Design Centre in Pune followed. Another collaboration, with Rajiv Raman and Dilys Thomas, produced A toolbox for fast and approximate solutions for large linear and semidefinite programs, published in the proceedings of the 2014 ACM India Computing Conference. It is a title unlikely to be mistaken for airport reading. It is also unusually clear about the ambition: solve large mathematical problems faster, using approximation.

In this setting, approximation has a precise purpose. An exact answer may demand more computation than a practical situation permits. Researchers study how to obtain useful answers with controlled compromises. The work involves the multiplicative-weights technique, a method for adjusting weights repeatedly as information accumulates. The important tension is between the quality of a solution and the resources required to reach it. That tension gives the early research a practical resonance well beyond a conference paper.

Learning to carry the answer into a business

Between academic research and Blackcoffer came industry experience. Bidyarthy worked in data science at InnovAccer, then joined the founding team of Gullak as co-founder and chief data scientist. His career chronology places those periods in 2013-2015 and 2015-2016 respectively. The sequence moves from investigating methods to applying them within organizations, and then to helping establish an organization himself.

His older professional writing has the vocabulary of someone comfortable with the tools: statistical modeling, predictive analytics, computational finance. His later public work adds the questions a business buyer asks. Can colleagues use the information? Can a system be maintained? Will a workflow survive contact with the people expected to run it? Those questions widen the job considerably. A mathematically interesting answer still has to find a home in someone’s working day.

Blackcoffer, founded in 2016, made software and analytics consulting the center of that work. Its services include data-driven dashboards, application development, information management, and big data analytics. Bidyarthy leads the company as founder and CEO. The firm’s operations span India and Malta, with a California presence as well. His professional setting is international, although Delhi remains the location associated with his own profile.

The five-terabyte opening

The 2018 infrastructure article is revealing because it starts with a limit. Below five terabytes, Bidyarthy’s published advice is to keep the setup modest. Make the data accessible through SQL, select a business intelligence tool, and avoid acquiring systems merely because a larger organization might need them. The recommendation gives the reader permission to postpone complexity, an underappreciated courtesy in a field with an abundant supply of products to buy.

“If you have less than 5TB of data, start small.”

Ajay Bidyarthy, Cloud Data Infrastructure, 2018

What makes the argument interesting is its audience. The point of accessible data is that people outside the engineering team can ask questions themselves. A dashboard is only one part of that arrangement. Underneath it lies an organizational choice about who gets to investigate the business. Making information easier to query can turn a request waiting in somebody’s queue into a question a colleague can pursue directly.

Read alongside the optimization research, the advice suggests a continuity of emphasis: resources matter, and usefulness depends on the problem being solved. The connection is an interpretation of the work, rather than a claim that a single algorithm explains an entire company. Still, the family resemblance is there. Both subjects ask when additional effort produces enough additional benefit to justify itself.

A different kind of network

Netclan brings a second meaning to the word network. Bidyarthy is its co-founder and CTO, working with founder and CEO Brijesh Awasthi. The platform seeks to connect people, professionals, and businesses through location and shared interests. The intended unit is a nearby community. Someone useful to you might be close enough to meet and still absent from every contact list you already possess.

The distinction between global reach and local relevance is central to that product. A professional network can contain people across continents while offering little help with an immediate neighborhood need. Netclan’s ambition is to make nearby people and opportunities discoverable. In its description of Bidyarthy, the company places family and friendship alongside his institutional and professional credentials. The order is a small human detail in a biography otherwise dense with technical nouns.

There is an intriguing pairing here. His early security research considered patterns inside computer networks. Netclan concerns discovery inside networks of people. The purposes are different, but each involves information that becomes more valuable when it can be placed in context. Proximity, like a pattern of behavior, helps determine which connection deserves attention.

A habit of leaving useful instructions

Bidyarthy’s writing extends beyond company descriptions. A June 2019 entry called Political Intelligence Database assembles routes into election finance information, voting data, and public records. It points readers toward resources including the Federal Election Commission, Voteview, and census material. The format resembles a working reference list. Its interest lies in showing where an inquiry can begin, rather than announcing the conclusion before the information arrives.

In March 2025, he published a list of data analysis tools organized by function. Programming languages sit alongside business intelligence products and other analytics categories. That arrangement makes a crowded technical market easier to navigate by beginning with the job a tool performs. In August 2025, posts under his byline covered question routing and a graph-based chatbot, including the use of external information tools to support responses.

The subjects have changed since the student research, but the instructional impulse is recognizable. A list sorts available tools; a workflow sorts possible next steps; a data reference sorts places to look. These are modest forms of explanation. They give another person something concrete to use, which is a reasonable test for a technical article and a considerably harder one than sounding knowledgeable.

The cook behind the computing

Outside the professional inventory, Bidyarthy says he enjoys travel, adventures, television series, movies, and politics. He also likes trying and cooking new dishes. That last detail offers a welcome interruption to the algorithms. A founder who works with mathematical models also spends some spare time making things whose success can be judged without opening a dashboard.

His stated professional ambition is to build technology and data-driven products with founders, scientists, statisticians, designers, and developers. The range of collaborators matters: the work he describes needs people who can formulate a question, build a system, and make it understandable. From Jamui and Kolkata to Guwahati, research labs, and entrepreneurship, his career has accumulated different ways of approaching that task. The small beginning remains a useful place to look: give someone access to the information, then help them do something with it.