First Commit to First Customer: How Startup Takton Shipped in Just Two Months with Tiger Data

A 5-person team designed their entire stack around Tiger Data before writing a line of production code. Two months later, devices were live on shop floors.

This is an installment of our "Community Member Spotlight" series, in which we invite our customers to share their work, spotlight their success, and inspire others with new ways to use technology to solve problems.

Farbod Moghaddam is CTO and co-founder of Sense Manufacturing (Sense), a startup building affordable machine monitoring for small and mid-size manufacturers. In this overview, Farbod shares why he selected Tiger Data before writing a single line of production code and how Tiger Data enabled his team to ship their first field devices in under 60 days.

Fast, Affordable Manufacturing with Takton

Daniel Scott Mitchell and Carson Morell founded Takton after years working closely with manufacturers and their supplier networks. Again and again, they saw the same problem: most industrial technology was designed for large enterprises, leaving small and mid-sized shops locked out by high costs, rigid device minimums, and complex deployment requirements. Mom-and-Pop manufacturers, the backbone of the supply chain, were often forced to operate without modern visibility tools. Takton was created to change that reality by building purpose-built products that meet real shop needs without compromise, but at price points smaller manufacturers can actually afford.

Sense emerged as Takton’s solution for machine monitoring. To execute the vision, Takton brought in Farbod Moghaddam to lead the IoT and firmware architecture, adding deep expertise in connected devices and edge systems. Together, the team set out to rethink how machine monitoring could be delivered to the shops that need it most. Sense devices monitor machine power consumption and vibration to determine machine status, securely streaming this data to the cloud every few seconds. Shop owners gain immediate visibility into machine uptime, receive alerts when vibration or power signatures shift, and can investigate anomalies before they escalate into costly failures.

The team also redesigned the business model around accessibility. In a market defined by 10-device minimums, Sense allows customers to start with a single device if needed. Where competitors typically require multi-year contracts, Sense offers simple month-to-month subscriptions with no penalties. The result is significant cost relief for smaller manufacturers, approximately 30% lower total cost in year one and up to 70% lower in subsequent years compared with the closest competitors, making enterprise-grade machine visibility attainable at shop scale.

Building Our Tech Stack from Scratch

The team initially evaluated InfluxDB as the time-series foundation for the Sense Manufacturing platform. While InfluxDB performed well for high-frequency data across a limited number of streams, the production requirements involved ingesting data from thousands of devices, each reporting power and vibration a few times per minute. InfluxDB couldn’t deliver on this use case for high ingestion across thousands of devices at scale. 

The team also wanted to avoid stack fragmentation. Their first product, CarbonReport, ran analytics on Supabase using standard SQL and Postgres. Adopting InfluxDB would have introduced a second query language and storage paradigm, which didn’t make sense for a small team to maintain.

Why We Chose Tiger Data as Our Time-Series Database

The path to Tiger Data started with a Reddit thread. While researching time-series database trade-offs comparing Prometheus, InfluxDB, and TimescaleDB, I came across an article by Mike Freedman, Tiger Data’s CTO, breaking down the performance characteristics of each stack. The article made the decision clear.

The main takeaway for me was: in our use case, where we're going to have thousands of devices in the field requiring real-time dashboard analytics, TimescaleDB is the best choice. - Farbod Moghaddam, CTO & Co-founder, Sense Manufacturing Inc

For a small number of high-frequency streams, InfluxDB works well. But once you’re ingesting high-rate data from thousands of devices, Tiger Data’s architecture is a better fit. Because it’s built on Postgres, we kept a single SQL-based stack across both products, so any developer could work in either codebase without switching context. 

We then had to decide whether to self-host TimescaleDB or use Tiger Data’s managed service, Tiger Cloud. After weighing the tradeoffs, the team determined that adding the operational burden of running a production database cluster, including backups, upgrades, failover, and storage monitoring, did not make sense for a two-engineer team. Instead, they chose Tiger Cloud to offload that infrastructure management. 

We built around Tiger Data. It was a very key piece - it was definitely not an afterthought in terms of the architectural design. - Farbod Moghaddam, CTO & Co-founder, Sense Manufacturing Inc

Inside the Sense Manufacturing Stack

The Sense architecture splits cleanly into four layers: the edge device, AWS-hosted ingestion services, external managed storage, and the customer-facing web portal.

At the edge, each Sense device runs a Linux-based firmware designed for reliable industrial deployment. The device measures machine power using amp clamps placed around single- or three-phase power lines and captures vibration through a three-axis accelerometer mounted to the machine body. These readings are transmitted every few seconds over Ethernet or Wi-Fi to an MQTT broker in AWS using TLS encryption. Firmware updates are delivered remotely through an OTA update service over the same secure connection.

The broker forwards each message to Takton’s edge device processing and routing service. This service validates the device against the Sense backend and loads the machine’s calibration thresholds, such as power and vibration boundaries that define idle versus running states. Incoming telemetry is then evaluated against these thresholds and the device’s current state. If the state changes—for example, from idle to running or running to alert—the service records an event in Tiger Cloud and pushes an update to the frontend. If the state remains the same, the raw telemetry is still persisted to Tiger Cloud, but no new event is emitted, keeping the event stream clean.

The storage model follows the shape of the data. A relational PostgreSQL database stores low-churn operational data such as device registrations, users, calibration settings, and customer organizations. Tiger Cloud stores the high-volume time-series telemetry, including power readings, three-axis vibration, and machine state events. With more than 100,000 machine-hours recorded, the automatic partitioning provided by TimescaleDB hypertables has been essential for maintaining fast and reliable data ingestion.

The customer-facing Sense Manufacturing web portal queries Tiger Cloud directly for the raw time-series data used in calibration views and sensor visualizations, while the Sense backend provides user, device, and configuration data. Commands such as configuration updates and threshold adjustments are routed back through the MQTT broker to devices deployed in the field.

Zero to Production in 60 Days

Development began in July 2025, and by September the first devices were already running in customer facilities. Two engineers built the firmware and web application in parallel. That speed was possible because the team deliberately chose technologies they already knew and relied on managed infrastructure where it made sense, allowing them to focus on product development instead of operational overhead.

Tiger Data also earned the team’s trust early. Even small support tickets received fast responses and quick resolution from the engineering team. That responsiveness gave the team confidence that Tiger Cloud could reliably support a production system.

Within two months, we went from nothing to the first devices being installed. - Farbod Moghaddam, CTO & Co-founder, Sense Manufacturing Inc

Immediate Value for Small Manufacturers

Once the product went live, small manufacturers immediately realized the benefit of our offerings. 

One pilot customer realized they could have avoided significant maintenance costs by activating Sense Manufacturing’s alerts. As outlined in the machine failure case study, the customer’s CNC machine ran out of lubricant and failed, resulting in a $50,000 repair and roughly three months of downtime, costing hundreds of thousands in lost production.

Afterward, they asked whether Sense had seen signs of the failure in advance. It had. The system had been flagging out-of-range vibration readings for four days prior to the failure, but notifications were not enabled at the time. The alert fired because Tiger Cloud retained the full-resolution vibration data, and the calibration logic correctly identified the anomaly. If that data had been averaged or aggressively downsampled, the signal likely would have been lost. The customer immediately enabled notifications to help prevent future failures that could be avoided with the Sense platform.

We were alerting for four days prior, but the customer just hadn't turned on our notifications. If they had, it would have been a $3,000 part change instead of the $50,000 rebuild. - Farbod Moghaddam, CTO & Co-founder, Sense Manufacturing Inc

A small CNC shop running a few dozen machines encountered a different kind of insight. The owner was surprised to see that one machine showed no recorded uptime for a particular day when it was expected to be running. His plant manager also believed the machine had been running, but Sense clarified the situation with data. The power consumption and vibration readings showed a pattern consistent with the machine being powered on throughout the day, but not one consistent with actively running jobs. The data revealed the root of the discrepancy: the machine had been turned on, but no work had been scheduled for it. Sense provided the missing context. By monitoring both power consumption and vibration, the system can distinguish between a machine that is powered on but idle and one that is actively producing parts.

Looking Ahead

The long-term roadmap for Sense includes advanced features designed to give manufacturers deeper insight into machine health and emerging issues. These capabilities rely on aggregating large volumes of machine telemetry over time. As power and vibration data accumulates in Tiger Cloud, it creates a historical baseline of normal machine behavior across machines and jobs. With a sufficiently large dataset, Sense plans to apply statistical and machine learning methods to identify subtle changes in vibration patterns and other signals, allowing potential failures to be detected earlier than would otherwise be possible. The time-series data being collected today forms the foundation for these future capabilities.

It's a ripple effect. You guys help us as startup founders, and then because you helped us out, we get to help out the small manufacturers." - Farbod Moghaddam, CTO & Co-founder, Sense Manufacturing Inc