My journey to founding Runlog has been shaped by deep technical expertise in AI/ML, proven execution at scale, and firsthand experience with the challenges of deploying AI systems. I'm building the infrastructure layer that enables teams to deploy trustworthy AI systems to production.
Runlog
Building Runlog Atlas, the infrastructure layer for human-in-the-loop AI systems. Drawing on ML expertise from Meta and startup experience from Bridge and Dirac, solving the critical problem of how teams can deploy trustworthy AI systems to production.
Runlog Atlas provides confidence-first design, priority queue review, and reusable human judgment—targeting the $10B market opportunity as AI systems move from demos to production.
Bridge
Built document intelligence pipelines processing 10,000+ documents daily. Watched teams review nearly every output despite 90%+ extraction accuracy — and founded Runlog Atlas to solve the review bottleneck that extraction tools never address.
Dirac - Head of AI
Head of AI at a CAD-automation startup. Processed 1M+ geometries and unstructured engineering PDFs; built confidence scoring and human-review workflows from scratch. Reduced user-facing latency by 70% and workflow interruptions by 90%.
Meta
Built ML systems serving 50M+ users daily. Improved hate organization detection by 15% (PR-AUC), directly impacting 11M+ profiles daily and reducing false negatives by 39%. Learned that confidence scores without operational routing are meaningless.
Stax - Founder
Founded Stax, growing to 400+ weekly active users across 4 colleges, supporting 15,000+ classes. Invested $10K and managed the entire product lifecycle. Learned invaluable lessons about product development, user acquisition, and market validation.
Georgia Tech
Earned degree in Mechanical Engineering with a minor in Computer Science (Intelligence thread), 3.4 GPA. Financed entire education through internships at Microsoft, Capital One, and Cardlytics while maintaining strong academic performance.
"I've spent years building AI systems that extraction teams trust. Every one of them hit the same wall: humans reviewing everything, judgment discarded after each review, teams drowning as volume grew. Atlas makes judgment reusable — so review effort shrinks as data scales."
— Vatsal Trivedi, Founder & CEO, Runlog
Technical Execution
- •AI/ML at Scale: Production ML systems serving 50M+ users, processing millions of items daily
- •Infrastructure: Built systems from 0→1, handling 10K+ docs/day and 1M+ geometries
- •Performance Optimization: 70% latency reduction, 90% workflow improvement, $300K annual savings
- •Full Stack: Python, TypeScript, React, Next.js, TensorFlow, PyTorch, AWS, GCP
Founder Qualities
- •0→1 Building: Founded Stax (400+ users), led multiple greenfield projects at Dirac
- •Resilience: Self-funded education from age 16, managed full lifecycle from idea to launch
- •Team Building: Currently mentoring 8 engineers, proven track record of cross-functional collaboration
- •Market Insight: Deep domain expertise in AI observability, lived the problem at Meta and Bridge
Product & Strategy
- •Product-Market Fit: Validated customer pain through direct experience building at scale
- •Strategic Thinking: Category-level insight (extraction vs. review), not feature thinking
- •Go-to-Market: Clear ICP (private markets ops), pull-through strategy (legal, research)
- •User Acquisition: Grew Stax to 400+ users across 4 colleges, enterprise sales experience
Domain Expertise
- •AI Observability: First-hand experience with review bottlenecks at Bridge (10K+ docs/day)
- •Document Intelligence: Built pipelines at Bridge, understand extraction vs. review dynamics
- •Enterprise AI: Deployed production ML at Meta (50M+ users), know what breaks at scale
- •System Architecture: Built from scratch at Dirac, understand reliability and performance requirements
Interested in Runlog Atlas, discussing AI infrastructure, or exploring collaboration opportunities?