Inside, Dr Chaitanya K. sits in front of a microscope and a half-drunk cup of tea. On the wall, a faded sign reads in hand-painted Marathi: “Accurate diagnosis, unwavering trust.”

But here, accuracy has always been a battle fought against exhaustion, distractions and overwhelming demand.

A paediatric blood report arrives on Chaitanya’s screen. The numbers seem ordinary: white blood cells normal; haemoglobin slightly low; platelets borderline. Typical monsoon fever, he thinks, almost ready to dismiss it. But experience and the scars of past misdiagnoses have taught him caution. Instead of signing off, he gently places the slide into an artificial intelligence (AI)-powered scanner he recently bought, against nearly everyone’s advice.

Amid the clamour rising from the street, the machine begins examining hundreds of cells, undistracted by the relentless noise below.

Then, a red notification flashes on the screen. Blast cells: 86%.

Chaitanya stares briefly. There’s no ambiguity.

“That’s leukaemia.”

Across town, in a dimly lit paediatric ward at the government hospital, Prasad Pawar, a sugarcane farmer from a village nearby, wipes sweat gently from his son’s forehead. The 12-year-old boy lies listless, IV fluids dripping slowly into his arm. For days they had moved from clinic to clinic, first an orthopaedist for back pain, then a paediatrician suspecting dengue. No one said the word “cancer.” No one thought it necessary to look deeper.

“He had some pain in his back, so we went to the orthopaedic,” Pawar told this writer later.

Earlier that afternoon, a pathologist Pawar had never met—someone who had failed class 12, someone who taught himself pathology from social media site X, someone who learned the hard way not to trust tired eyes alone—had seen something no one else had.

Minutes after the machine flagged the anomaly, Chaitanya called Pawar’s paediatrician.

“Refer them immediately,” he said calmly.“Sambhajinagar, now.”

There was no debate. The boy’s name didn’t matter yet, nor did Chaitanya’s past failures. All that mattered were the numbers the machine refused to overlook—and the life they might still have a chance to save.

In India’s small-town labs, accurate diagnoses are often missed—fatigue, human error and limited resources turn routine tests into silent gambles. But AI is changing the rules. This isn’t Silicon Valley’s grand vision of AI replacing doctors. Instead, it’s a story about overlooked places like Parbhani, where technology helps exhausted pathologists see clearly again, catching life-threatening diseases before it’s too late. In India’s small-town labs, accurate diagnoses are often missed—fatigue, human error, and limited resources turn routine tests into silent gambles. But AI is changing the rules.

Lessons from XIn the margins of Chaitanya’s résumé, beyond medical degrees and equipment manuals, lies a quieter credential: failure. Not the motivational kind of failure that fills TED talks and commencement speeches, but the stark, personal kind that nearly ends a journey before it begins.

Chaitanya grew up in Parbhani, a small district town in Maharashtra better known for sugarcane fields than second chances. His father, an electrical engineer, died in an accident when Chaitanya was in school, leaving him to be raised by his mother. In class 12, the crucial Indian exam that determines academic futures, he failed.

“I failed class 12… Everyone told me to give up,” he recollected.

But his mother refused to accept defeat, and slowly, so did he. Over the next three years, he regrouped, clawed through self-doubt, and finally gained admission to a medical college. His original dream—neurosurgery, inspired by his father’s early death—was out of reach. Instead, he found himself in pathology, not by choice but circumstance: his exam scores weren’t high enough for a competitive surgical residency.

At first, pathology felt like a compromise. But slowly, it turned into something deeper. In 2018, while teaching pathology at Smt. Kashibai Navale Medical College and General Hospital in Pune, Chaitanya stumbled upon a Leica digital scanner—unplugged and forgotten, gathering dust in a corner of the department.

He asked around; no one seemed to care if he took it. He contacted the manufacturer for training and began experimenting, scanning pathology slides digitally and sharing images on X. What started as idle curiosity soon became his education. Pathologists from across the world responded to his blurry scans, corrected him, debated diagnoses, taught him what textbooks couldn’t. ‘X changed my entire scenario. Instagram reels don’t teach you pathology.’ — Dr Chaitanya Khillare

“X changed my entire scenario,” he said. “Instagram reels don’t teach you pathology.”

Each night, after his wife and kids went to sleep, he spent hours analysing blood smears, studying subtle variations, understanding how even trained doctors could disagree over what constituted a “blast” cell—the marker of leukaemia. Most importantly, he reflected on his own diagnostic mistakes.

“I’ve missed cancer myself,” he said. “Twice.”

He said it plainly, without drama—the confession of someone who had spent countless hours hunched over a microscope, certain he had seen enough, only to realize later he hadn’t. He missed diagnoses not out of carelessness, but because he was exhausted. This was why the AI scanner mattered. Not because it was infallible, but because it never grew tired, never skipped the 10th slide out of fatigue.

“AI isn’t about speed,” Chaitanya said. “It’s about not failing someone just because we’re exhausted. Or because a pattern didn’t jump out the first time.”

The Leica scanner had given him his first glimpse of technology’s potential. But the urgency crystallized only when he returned to Parbhani to open his own pathology lab above a cardiology clinic. Running a 24-hour operation, he faced a daily reality of exhaustion and relentless demand—blood reports needed immediately, cardiac emergencies arriving at all hours. Fatigue was no longer theoretical; it was inescapable.

When he heard about SigTuple’s AI-powered device—one capable of scanning blood smears, tagging cells and detecting abnormalities—he didn’t ask if it would be profitable. He asked if it would see what he sometimes missed.

SigTuple, a Bengaluru-based AI startup specializing in medical diagnostics, builds tools designed to catch diseases before it’s too late.

Against everyone’s advice, Chaitanya bought it—a ₹6.5 lakh investment in a town where most labs still used manual counting.

What the machine sawIt was the SigTuple AI100 scanner that caught Prasad Pawar’s son’s leukaemia.

The device methodically analysed the smear, finding and focusing on the monolayer, where cells lie evenly spaced and clearly defined. Within minutes, it scanned hundreds of cells, measuring their size, shape and patterns, a task nearly impossible for an exhausted pathologist reviewing dozens of slides a day.

No one had explained algorithms or AI to Prasad. But what he understood instinctively was that something had looked closer—exactly when others had looked away.

Within a day, the diagnosis was confirmed through flow cytometry at Sambhajinagar, a city with better medical facilities, about 200km from Parbhani. Treatment began immediately. The boy remained stable, under careful monitoring.

For Chaitanya, such outcomes reaffirmed his faith in the partnership he’d forged with his machine—one that wasn’t always about spotting disease, but sometimes about ruling it out.

Eyes that don’t tireChaitanya didn’t buy the scanner because of a glossy pitch or hype. In fact, there was no pitch at all—just a promise from a machine that didn’t care whether the slide was stained perfectly or the day had run too long. He trusted it because it caught things he sometimes missed. And that trust was built on years of meticulous engineering by someone who understood how fragile frontline diagnosis could be.

The AI-powered scanner that flagged 86% blast cells in that boy’s smear was developed by Tathagato Rai Dastidar, co-founder of SigTuple. Tathagato had spent years building tools meant for labs that never appeared in glossy brochures—places like Parbhani, where blood samples are mis-processed, staining is uneven and technicians might never have completed formal training.

“You walk into a lab in India, and 8 out of 10 times, the smearing or staining is poor,” Dastidar told this writer. “If we didn’t build for that, we’d be building for nobody.”

View Full Image

Tathagato Rai Dastidar, co-founder of SigTuple.

SigTuple’s scanner doesn’t simply magnify slides; it replicates the thought process of a skilled pathologist at machine speed. A single drop of blood, smeared on a glass slide and stained, is loaded into the AI100. The scanner quickly finds the monolayer—the ideal zone where cells are clear, distinct and undistorted. It then captures high-resolution images, identifies and counts white blood cells, and pinpoints any morphological anomalies.

It doesn’t stop at a glance. It reviews dozens, sometimes hundreds, of cell images to ensure nothing subtle is missed. In under two minutes, it performs a level of analysis that would take a human nearly half an hour, flagging cells that appear suspicious, uploading findings to the cloud and creating a detailed digital report.

“We’ve seen staining quality improve in labs within days of installing our scanner,” Dastidar said. “Because now, nothing gets discarded. Everything is seen.”

What SigTuple created is more than a faster microscope; it’s a workflow that remembers. Once analysed, no smear vanishes into a forgotten folder. Each questionable cell, each subtle anomaly, can be retrieved, reconsidered or re-examined months or even years later.

“This isn’t about man versus machine,” Dastidar explained. “AI sees the sample. It doesn’t know the patient. But when used correctly, it helps the doctor see better.” ‘AI sees the sample. It doesn’t know the patient. But when used correctly, it helps the doctor see better.’ — Tathagato Dastidar

This subtle yet crucial partnership bet-ween human intuition and machine vigilance is precisely what Umakant Soni, an experienced AI ecosystem builder closely familiar with SigTuple, emphasized.

“India’s rural and semiurban healthcare infrastructure is fragmented, stretched thin, often running under conditions of constant fatigue,” Soni said. “You see compounders playing the role of doctors. Technicians working without proper training. Historical patient data is rare or non-existent.”

For Soni, the critical insight behind SigTuple’s technology wasn’t simply its sophistication under ideal conditions—but its resilience under real-world pressures. “AI models here must adapt to noise, to incomplete information, to messy realities,” he said. “They must work offline, speak in vernacular languages, and fit into the chaotic rhythms of frontline care.”

Soni believes systems like SigTuple’s can dramatically reshape rural healthcare. “There’s a powerful case for using AI-driven screening at the primary health centre level,” he argued. “Right now, too many preventable cases escalate into full-blown health crises that overwhelm urban hospitals like Tata Memorial. Early diagnostics could stop these tragedies before they start.”

A different kind of missNot every misdiagnosis unfolds in a rural lab or culminates dramatically in a paediatric leukaemia ward. Some happen over days and weeks, even in cities where good doctors and advanced diagnostics are easily accessible. But the experience—the confusion between what you’re told and what you feel—is no less disorienting. That’s why Anuruddh Mishra’s story belongs here, as a counterpoint to the urgency that saved a child in Parbhani.

View Full Image

A file photo of Anuruddh Mishra. (August.AI website)

In early 2020, just weeks before India entered its covid-19 lockdown, Anuruddh, a tech product builder living in Bengaluru, began noticing stiffness in his fingers. Then came sharp pains in his knees. Initially, he brushed it off. But as discomfort grew, anxiety prompted what most urban Indians with access tend to do: he booked multiple tests, consulted telehealth platforms and tried to decode medical jargon on his own.

Three different labs returned the same troubling result: elevated anti-cyclic citrullinated peptide (anti-CCP) levels—a biomarker that pointed toward early-onset rheumatoid arthritis. A specialist agreed it was worrying, though perhaps not advanced enough yet to warrant aggressive medication. Uncertainty lingered.

Over the months that followed, his pain intensified. One morning, as he stood cooking breakfast, Anuruddh’s right knee abruptly gave out beneath him. “I told myself I’d walk into the airport that day without assistance,” he later recalled, still vividly remembering his determination to resist the narrative of disability creeping into his twenties.

Then came an unexpected twist. New tests from another lab returned clean: no anti-CCP. No rheumatoid arthritis. The source of his discomfort turned out to be something entirely different—a simple dietary imbalance involving elevated uric acid and vitamin D deficiency, easily manageable with changes in diet and supplements.

Relief, however, was quickly replaced by frustration. “I had access. I wasn’t poor, wasn’t rural,” Anuruddh explained, visibly animated by lingering disbelief. “But for months, I was left guessing, worrying and confused. All I had were conflicting numbers.”

This personal turmoil sparked the creation of August.AI, a conversational assistant designed not to replace doctors, but to offer clarity and reassurance in the chaotic silence between visits. “I wasn’t building an AI doctor,” Anuruddh emphasized. “I was creating the companion I desperately needed—one who says, ‘I see exactly what you’re seeing. Let’s figure this out together.’”

View Full Image

A screen grab from August.AI’s website.

Today, August.AI serves millions of patients, primarily in towns far smaller and less equipped than Bengaluru. It has prevented dangerous drug interactions, alerted liver donors to contraindicated medications and correctly identified the cause of fainting episodes as a missed vaccine. Anuruddh’s own months-long brush with confusion turned into a deeply empathetic digital companion.

He isn’t a doctor, but his experience echoes the central question that Chaitanya had grappled with in his own lab: What if someone had just looked a little closer?

The lab that doesn’t forgetMost medical labs in towns like Parbhani aren’t built to remember—they’re built to churn. Samples arrive, are processed quickly, and the paper reports vanish into the vast machinery of India’s fragmented healthcare system. Patients rarely return for follow-ups, and their medical history often fades, lost or discarded.

But Chaitanya’s lab refuses to forget.

“Patients vanish,” he said, voice steady yet gentle. “But their cells don’t have to.”

This isn’t merely a line he repeats; it’s the core philosophy of the lab. Here, each smear and scan is uploaded, meticulously archived and stored securely in the cloud within minutes, even at odd hours when most technicians are sleep-deprived and mistakes are commonplace. The lab runs 24 hours, reflecting the urgent heartbeat of the cardiology clinic below. Night or day, emergencies arrive. Slides are hurriedly prepared, fed into the AI scanner and remembered permanently.

View Full Image

Dr Chaitanya Khillare with his colleague Dr Shivkanya Rasal, outside their diagnostics lab in Parbhani.

Memory, Chaitanya has learned, can be transformative. One case stood out vividly when we spoke: a young girl whose fever had lingered too long. Her paediatrician suspected rheumatic fever—a condition easily missed but devastating in its consequences if not diagnosed early. The initial lab results—an anti-streptolysin O (ASO) titer—came back normal. But Chaitanya couldn’t shake his unease.

“The doctor kept saying it didn’t feel resolved,” he recalled. “I felt it too.”

Chaitanya insisted on another test, sending the sample nearly 300km away to a certified lab in Hyderabad. Three tense days later, the results came back positive, confirming the initial suspicion—acute glomerulonephritis, a serious kidney condition.

“That night was restless,” Chaitanya confessed. “She was just a child. If that report had been missed, six years later, she would surely have developed heart disease.”

Chaitanya’s refusal to erase ambiguity can feel counterintuitive in a system that values quick results and neat answers. Yet it has already proven crucial. Once, a young girl arrived with a swollen lymph node. Lymphoma was suspected, but her family, wary and perhaps overwhelmed, refused further tests. They disappeared without leaving even a phone number behind.

Almost a year later, the girl returned. This time, the mass was larger, visibly worse. Chaitanya, recalling their abrupt departure, didn’t have to rely on fragmented memory or scribbled notes. He still had the cells—from the very first visit, carefully digitized and saved.

“I didn’t need a fresh biopsy,” he explained. “I had the truth, right there, from the last time she came.”

Diagnostics, especially in places like Parbhani, is often about speed, about navigating between fatigue and efficiency, about clearing the next report from a backlog. But for Chaitanya, persistence has become just as crucial as precision. He sees his AI scanner not only as a diagnostic companion but as a witness, capable of remembering every overlooked cell, every ambiguous smear, every patient who slips from the waiting room into uncertainty.

View Full Image

Dr Chaitanya Khillare with his colleagues at the lab.

Not to fail someoneIn Parbhani, where diagnostic delays often mean irreversible outcomes, there is little room for error.

“It’s not about being faster,” Chaitanya told this writer late one evening. “It’s about not failing someone because we were tired or distracted.”

This commitment connects him to others who refuse to accept the routine failures in healthcare.

Tathagato Rai Dastidar, who built SigTuple’s scanner, didn’t create it for venture capital pitches or tech headlines. He built it to withstand the frustrating realities of Indian labs—poor staining, uneven training and weary technicians. “AI doesn’t blink,” Dastidar said. “Every sample gets the same attention.” In Bengaluru, Anuruddh Mishra felt first hand the pain of misdiagnosis. It drove him to build August.AI—not to replace doctors, but to offer clarity when medical answers felt cold or distant. These three paths cross not because they share a vision of futuristic technology, but because they refuse to normalize neglect.

Not every story ends well. Chaitanya knows this. Weeks after this writer’s visit, he described a case that still haunted him. Late one night, a three-year-old’s blood report showed extremely high white blood cell counts. The AI quickly ruled out leukaemia and sepsis. Troubled, Chaitanya called the paediatrician and urged him to look deeper.

Hours later, the doctor discovered the child had stopped drinking water and was showing signs of hydrophobia—a hallmark of rabies. There was no visible bite or wound history, just the silent, devastating progression of the disease. The child, born after his parents tried for 17 years, could not be saved.

“AI didn’t diagnose rabies,” Chaitanya later wrote on LinkedIn, openly grappling with the outcome. “But it helped rule out leukaemia and sepsis quickly. Sometimes integrity means doing everything right—even when it won’t change the outcome.”

For Chaitanya in Parbhani, AI isn’t revolutionary. It’s just one more safeguard against the simplest yet harshest failure: missing something because you were too tired, too rushed, or too human.