Every afternoon, Satish Kumar had two and a half hours with somebody else's physics notes. The class was beyond his family's means. A friend of his brother took it, came home, handed over the notebook, and waited while Kumar copied what he needed. Then the original went back. It was an improvised supply chain for knowledge - narrow, time-boxed, and entirely dependent on another person's generosity.
Kumar grew up in a poor family in rural northern India. His mother died when he was young. His elder sister, whom he remembers as smart and capable, set aside her own ambitions to care for the family. He saw ability distributed widely and opportunity rationed by money, circumstance, and network. The lesson did not arrive as a slogan. It arrived as borrowed paper.
He reached the Indian Institute of Technology Kanpur and graduated in 1999, then continued to UC Berkeley on a fellowship, studying Management of Technology at the Haas School of Business. Those institutions matter to his biography, but his later work keeps circling a more uncomfortable idea: institutions are poor substitutes for evidence. A school can open a door. It cannot, by itself, prove what someone can do on the other side.
01 / The repeated questionCan you show the work?
Kumar's career began in engineering and management at Oracle, where he worked on RFID and mobility products for supply chains. He later co-founded GetVendors, an online marketing company acquired by MatchPoint. Then came education technology: Snapwiz built digital learning and assessment products used by publishers including Wiley and McGraw Hill, while Edulastic brought assessment into classrooms. Edulastic was ultimately acquired by Sumeru Equity Partners.
Across those changes - physical goods, learning, classrooms, labor - the same product question kept resurfacing. How do you turn a claim into something observable? In education, a student can watch a lecture without learning. In hiring, a candidate can possess an elegant résumé without being ready for the task. Activity is easy to log. Capability takes better design to reveal.
Glider AI started in 2020 because Kumar's own team felt the cost of getting that design wrong. Recruiters would search, screen, and pass a candidate along. A hiring manager would enter late, only to discover the person could not perform the work expected in the role. Both groups were working hard. The signal between them was weak. Kumar called the wasted expert time an “interview tax.” His first-principles fix was to move proof upstream.
From claim to decision
The product that followed tries to simulate the first day on the job. A developer codes in an interactive environment. A nontechnical candidate works through a role-specific situation. Interviews can happen live, asynchronously, or through a conversational AI system. The purpose is less to ask whether the candidate knows the right vocabulary than to watch how the candidate handles the work.
02 / Proof needs trustThe remote-hiring paradox
Remote work widened the talent map. Kumar described it as hiring “without borders,” a change that could place candidates outside familiar networks into serious consideration. It also widened the opportunities for deception. Glider reported a 92 percent increase in fraudulent assessment activity during the pandemic. More recently, generative AI has made polished answers, code, audio, and even synthetic identities easier to produce.
That creates a paradox. A skills test is fairer only if employers can trust who completed it. A verification system is fair only if candidates know how they are being monitored. Kumar has been explicit that Glider's remote assessments require candidate permission and should not operate behind the scenes. The software combines skill validation with identity checks and transparent proctoring. In his telling, integrity and competency belong in the same workflow because neither is useful without the other.
This is where his social mission meets a practical constraint. Removing a famous school from the top of the funnel can reduce one kind of bias. It does not guarantee a good decision. The replacement signal has to be relevant, consistent, and trustworthy. Kumar's preferred answer is a short, realistic assessment calibrated to the actual role, not a generic ordeal that exhausts the candidate.
03 / Access becomes architectureA personal story, made operational
Kumar's argument for competency above credentials is unusually personal for enterprise software. He has spoken about capable people from his childhood who left education because money or family duty intervened. He has also seen the same pattern inside his own company. Praveen, a Glider employee without the traditional corporate background for his eventual role, joined as an individual contributor and grew into responsibility for global customer support from India. Kumar remembers his ownership, customer instinct, and willingness to solve problems beyond the formal edges of the job.
The management lesson mirrors the product: do not let a proxy do the seeing for you. Kumar looks for what he calls an ownership gene. He values people shaped by difficult startups because failure can leave behind practical resourcefulness, provided dedication was real. His advice to CEOs includes building a deliberate delegation process, giving people a voice, and making strong operators co-owners of the journey.
There is levity in that advice, too. He tells leaders to grant their teams formal permission to have fun and to retain enough humility to laugh at themselves. His reading recommendation for entrepreneurs is Peter Thiel's Zero to One. Asked whom he would choose for a private meal, he named Oprah Winfrey, pointing to a self-made career and a public association with inclusion. These details do not explain a company, but they soften the machinery around it: a CEO who talks about metrics also talks about gratitude, voice, and the loneliness of decisions that affect other people.
One idea, four chapters
After IIT Kanpur and UC Berkeley, Kumar begins a long Oracle chapter in engineering and product management.
Education technology turns assessment into a product problem through Snapwiz and Edulastic.
Glider AI applies competency-based assessment to hiring, where weak signals create expensive interviews and missed candidates.
Findem announces an agreement to acquire Glider, pairing talent discovery with skills and identity verification.
04 / The next signalWhen AI writes the résumé
In 2025, Kumar introduced Agentic AI Interviews, designed to conduct a responsive conversation, test skills with hands-on tasks, and answer candidate questions across roles and languages. He then put himself through the system unscripted and unprepared. The demo was theatrical in a useful way: the person selling the examiner agreed to be examined.
Glider also launched an AI Assistant inside technical assessments and live coding interviews. The shift matters. If working with AI is becoming part of the job, banning it from every test may measure an obsolete form of isolation. Kumar wants employers to observe how a candidate collaborates with AI - where they reason independently, where they ask for help, and how they use the answer. The tool that complicates trust can also become part of the skill being tested.
The company reached several markers along the way. Staffing Industry Analysts named Glider the most innovative staffing technology in its 2021 Shark Tank competition. In March 2023, Glider announced a $10 million Series A from Primera Capital and other investors. Three years later, Findem announced a definitive agreement to acquire the company. The planned combination joins Findem's talent discovery with Glider's assessments, autonomous interviews, and identity verification. Terms were not disclosed, and the announcement said closing remained subject to customary conditions.
Kumar described the deal as a move from software toward outcomes: candidates discovered, validated, and verified before a hiring team sees them. It is a logical extension of the interview-tax idea. Make each transition carry evidence. Let recruiters spend less time reconstructing competence from claims. Reserve human attention for motivation, judgment, team context, and the deeper questions a simulation cannot settle.
05 / The useful theftReplace the proxy, then prove the replacement
There is a portable product lesson in Kumar's story. First, listen to the customer problem and read between the lines, as he advises founders. Second, identify the proxy hiding inside the workflow. A résumé stands in for ability. Course attendance stands in for learning. Years of experience stand in for proficiency. Third, build a small moment where the real thing can appear. Let someone perform a task. Make the evidence comparable. Keep the burden proportionate.
Then test the new signal with the same skepticism applied to the old one. Skills tests can encode bad assumptions. AI interviews can standardize a flawed rubric. Identity checks can become intrusive. Kumar's stated safeguards - job relevance, transparency, candidate permission, and human judgment after validation - are important because a fair-sounding process still needs inspection.
The borrowed notebook sits at the beginning of this story because it contains the whole tension. Kumar had the hunger and the ability. What he lacked was ordinary access. A friend, a fellowship, and a family's sacrifice helped bridge the gap. Software cannot reproduce that care. It can, however, stop treating access itself as proof of merit. Kumar has spent a career turning that distinction into assessment systems. The 2026 Findem agreement gives the idea a larger distribution path. The stubborn belief remains the same: when the door opens, let the work speak.