Best Place To BuildThe $45M Industrial AI Revolution | Daniel Raj David, CEO of Detect Technologies on BP2B S2 Ep.2
CHAPTERS
Detect Technologies in one sentence: actionable AI for safer industrial sites
Daniel opens with Detect’s mission: making the industrial world safer through “actionable” AI, not just analytics. The conversation sets up the core idea—AI that translates messy, multi-source industrial data into interventions frontline teams and leadership can act on.
What Detect actually does: one AI layer across cameras, sensors, and enterprise data
Daniel explains Detect’s product as a centralized AI layer that ingests visual feeds, time-series sensor data, and even ERP/notes data to produce real-time safety and operational insights. The system is designed to scale from small yards to large plants and report both incidents and culture/leading indicators.
Why safety is hard at scale: OSHA rules vs. limited safety teams
They discuss why traditional safety approaches fail: safety teams are small and cannot monitor every location continuously. Daniel cites global fatality numbers to frame the urgency and explains how safety standards (e.g., OSHA) define leading indicators that are difficult to track manually.
From detection to intervention: stopping machines and preventing fatalities
A key differentiation is that Detect doesn’t just identify unsafe situations—it can trigger real-world interventions. Daniel uses port crane operations as an example: detecting people under suspended loads and integrating back to machinery to stop operations and raise alarms.
Beyond cameras: predictive maintenance and IITM-born sensing tech (GUMPS)
The discussion expands into sensors and predictive maintenance, including high-temperature ultrasonic inspection challenges. Daniel describes a patented air-coupled sensing approach (magnetostriction) that works at high temperatures and requires AI to interpret complex signals to predict failures.
The data moat problem: training AI on rare hazardous events
Amrut asks how you get training data for hazards you want to avoid. Daniel explains that real-world data is crucial (synthetic data wasn’t viable early on), and that early industrial partners enabled access to the scenarios and labeling needed to build robust models—forming Detect’s long-term moat.
What ‘OSHA rules’ look like in practice: PPE, line-of-fire, vehicles, work-at-height
Daniel gives a practical taxonomy of safety categories Detect models, from simple PPE compliance to multi-factor probabilistic risk detection. They discuss how violations create human tragedy, legal liability, and site-wide morale impact.
Origin story at IIT Madras: one email, CNDE lab, and meeting Tarun Mishra
Daniel narrates how he moved from student introspection to hands-on industrial tech by reaching out to Prof. Krishnan Balasubramanian. He’s introduced to a novel sensor project, meets Tarun Mishra (industry practitioner and IITM alum), and the collaboration becomes the seed of Detect.
Student-builder life: CFI talent, industry trials, and running a ‘mini org’ before funding
Daniel describes spending much of his later IIT years on industrial pilots rather than campus life alone, while still engaging in sports and culture. Detect functioned like a student-driven organization with dozens of contributors, pulling in CFI talent to build early engineering capability.
Incorporation decision (2016): make-or-break as talent graduated and scope exploded
Detect incorporated in 2016 to retain talent and move from projects to a company, as multiple initiatives ran in parallel (sensors, drones, analytics). The team faced the classic inflection point: placements vs. committing to a startup, with mentors helping on company-building basics.
Pivots and the pandemic: from hardware+software to hardware-agnostic AI SaaS
They walk through Detect’s evolution: initially building proprietary hardware plus software, then realizing where the value truly sits. COVID disrupted on-site installs and drone services, prompting a major strategic pivot—no layoffs, retraining roles, and focusing on scalable SaaS over existing infrastructure.
Deployment architecture and scaling complexity: cloud, on-prem, and edge constraints
Amrut probes the technical delivery model. Daniel explains Detect supports cloud and on-prem deployments, plus limited edge processing when connectivity is poor—yet full OSHA-scale complexity often requires cloud compute due to rule/model breadth.
Becoming global (2019–2020): Shell as catalyst and the customer success bar abroad
Detect’s first major international contract (Shell) helped validate that safety is a global problem. Daniel highlights that global scaling is less about whether the tech works and more about meeting higher standards of implementation, support, and customer success—plus setting up regional entities.
Industry frameworks: Industry 4.0 (and beyond) plus how OSHA keeps evolving
Daniel breaks down Industry 1.0 through 4.0, positioning today’s AI/IoT/data-driven decision-making as the current wave, with 5.0 emerging in discourse. He notes OSHA and related bodies continuously evolve as new machinery introduces new risks, expanding rulebooks and investigation practices.
Funding, validation, and lessons for founders: obsession, co-founders, and ecosystem leverage
Daniel outlines Detect’s fundraising timeline from early angel/deep-tech seed to later global investors and strategic capital from Shell Ventures—plus customer awards as validation. They close with advice: founders must be obsessed, resilient to naysayers, supported by complementary co-founders, and active in leveraging IITM’s labs, mentors, alumni, and CFI talent.
Get more out of YouTube videos.
High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.
Add to Chrome