The global sleep-tech devices market is projected to grow from $25.26 billion in 2025 to $29.62 billion in 2026. Against that backdrop, a UK data engineer with experience in fintech, payments, and large-scale software infrastructure is building toward a category where AI does more than monitor sleep: it actively changes the sleep environment.
In this feature, we speak with Anand Rawat, an AI and data engineering leader whose latest project takes machine learning somewhere unusual: inside the mattress itself. With the global sleep-tech market projected to grow from $25.26 billion to $29.62 billion within a year, we wanted to understand what happens when a decade of banking and logistics infrastructure experience gets applied to something as personal as how people sleep.
Most sleep technology still stops at observation: it tracks movement, heart rate, or rest patterns, then presents the user with a report the following morning. Anand Rawat is working on a more active proposition. Through AINAP Ltd, the AI and data engineering leader is developing a smart-mattress platform designed to interpret live sensor data and respond while a person sleeps, using embedded devices, edge computing, and reinforcement learning to adjust pressure in real time rather than simply record what happened overnight.
What Happens When You Put Machine Learning Inside a Mattress?
Since April 2025, Rawat has been co-founder and majority shareholder of AINAP Ltd, where he leads the end-to-end technical design of an AI-native preventive-wellness platform centred on a working smart-mattress prototype.
The system uses ESP32 embedded devices connected through secure, low-latency MQTT pipelines to collect sensor data at the edge. That information feeds into a sleep-analytics engine capable of classifying posture and identifying behavioural patterns from live sensor inputs.
Its most distinctive element is an adaptive pressure-control system. Rawat is developing reinforcement-learning models designed to modulate mattress pressure in real time in response to sensor signals. Rather than limiting AI to passive sleep tracking, the platform is intended to make immediate adjustments based on how an individual is positioned and moving throughout the night.
AINAP remains pre-MVP, with its architecture currently at the working-prototype stage. Anand has already secured a partner and investor in Sarva Foam UK Ltd, the UK arm of Sarva Foam Industries, a polyurethane-foam recycling and manufacturing group founded in India in 2011 that processes around 15,000 metric tonnes of PU foam waste annually into rebonded foam used in products such as mattress cores and carpet underlays. Sarva Foam UK was incorporated in August 2024 as part of the group’s international expansion and is active in the UK carpet-underlay and floor-covering sector. The partnership includes a £5,000 initial seed commitment towards MVP development, with a further £250,000 in post-MVP investment contingent on delivery.
AINAP’s corporate structure, including Rawat’s majority shareholding, is recorded at Companies House.
Why Is a Data Engineer Building Hardware?
Rawat’s route into hardware is rooted in nearly a decade of software and data-infrastructure work. Between April 2022 and January 2023, he led a large-scale logistics simulation systems at Amazon UK, where he improved delivery-prediction accuracy by 18 percent and accelerated simulation experimentation cycles threefold.
At DataGardener, he developed credit-risk prediction models, synthetic-data generation pipelines using CTGAN, and ESG analytics systems that are now embedded in live financial and compliance workflows.
Earlier in his career, at Tata Consultancy Services, Rawat built backend systems for telecom platforms processing more than 10 million daily transactions. His work reduced incident-response time by 60 percent.
The common thread across banking, logistics, telecoms, and consumer wellness is the same: building systems that can absorb real-time data, interpret it quickly, and trigger a useful response rather than waiting for a manual review or retrospective report.
Is There Actually a Market for This?
The broader market suggests that the opportunity extends beyond conventional sleep tracking. The global sleep-tech devices market, including smart mattresses and wearable trackers, is projected to expand from $25.26 billion to $29.62 billion within a year, representing a 17.2 percent compound annual growth rate, according to Research and Markets, which forecasts the sector could reach $57.47 billion within five years.
Much of this growth is tied to a shift from passive monitoring tools toward systems that can adapt the sleep environment in real time. Wearables can capture data after the fact; AINAP is being designed to use that information immediately, with a modular architecture spanning hardware, firmware, and cloud-based AI layers.
Alongside his work with AINAP, Rawat remains a Technical Board Member at DataGardener, where he built B2BSorted, an AI-powered lead-discovery platform now deployed across more than 100 UK organisations and recognised through the AI FinTech100 and ESG Insight Awards.
It’s a pattern that shows up across Rawat’s work: take a system built for passive observation and give it a way to act on what it sees. B2BSorted did that for lead generation, turning a static company database into something that scores and surfaces opportunity in real time. AINAP is the same instinct applied somewhere far more intimate. If the reinforcement-learning models hold up outside the lab, the mattress stops being something that just reports on how you slept and becomes something that changes how you sleep, while you’re still asleep.
About Anand Rawat
Anand Rawat is an AI and data engineering leader with more than nine years of experience building production-grade systems across fintech, payments, B2B intelligence and healthtech. He is the co-founder of AINAP Ltd and a Technical Board Member at DataGardener. Rawat holds an MSc in Big Data Science from Queen Mary University of London.
