Saksham Goyal — AI Engineer, Building in Stealth

Code. Debug. Deploy. Measure. Repeat.

Shipping AI at scale. Even before the LLM hype.

Built end to end. AI models, backend, frontend, product.

Work

Stealth Builder

2025 — Present

Working on something at the intersection of AI and human intervention.Building from scratch. Product, engineering, go-to-market. More to share soon.

Head of AI and EngineeringIOK Labs

2021 — 2025

Software Engineer, AINayan Technologies

2020 — 2021

AI-powered traffic monitoring and road safety.Deployed violation detection for Dubai Police through Dubai Future Accelerators.

AI Research InternAdventum

2020

Medical imaging AI. Piloted with ophthalmology clinics in Spain.Built a pipeline to detect early-stage eye diseases from retinal scans.

Capabilities

Full-Stack Builder

Started with AI research. Built the backend to serve the models. Learned cloud to deploy them. Picked up everything else along the way because that's what building a product takes.

AI in Production

From detecting eye diseases in retinal scans to building multi-agent LLM systems that serve real users. Six years of shipping AI, not just training models.

Zero to One Product

No PM, no product team. Scoped features, ran sprints, prioritized roadmaps, and demoed to partners. Startups taught me product the hard way.

Growth and Revenue

Worked on brand partnerships. Captured partner requirements. Tracked revenue numbers. Built the tools to measure it all. At a startup, the business side is everyone's job.

Activity

Active GitHub contributor. View contributions at github.com/isakshamgoyal

Latest Writing

Anthropic Leaked Claude Code

A 59.8 MB debug file shipped with Claude Code v2.1.88. Inside: 512K lines of TypeScript, an unreleased model roadmap, and a Tamagotchi.

AlphaEvolve: A Step Towards Self-Evolving AI Agents

Google DeepMind's evolutionary coding agent uses Gemini to generate, test, and evolve code, breaking a 56-year-old math record and optimizing Google's own infrastructure.

How ChatGPT Agent Actually Clicks, Codes and Delivers

OpenAI's ChatGPT Agent combines a planning model, a computer-using agent, and a secure VM to execute multi-step tasks across the web.

Contact

Building something? Let's talk.. Happy to talk AI, tech, product, and everything in between. Grab a slot below or just drop a message.

Email: isakshamgoyal@gmail.comBook a Meeting: Schedule a callLinkedIn: isakshamgoyalGitHub: isakshamgoyalInstagram: isakshamgoyal
In stealth

Saksham Goyal

 

Building something in AI.

Code. Debug. Deploy. Measure. Repeat.

Shipping AI at scale. Even before the LLM hype.

Built end to end. AI models, backend, frontend, product.

Work

What I've built.

Stealth Builder

2025 — Present

Working on something at the intersection of AI and human intervention.
Building from scratch. Product, engineering, go-to-market. More to share soon.

Interested in investing or collaborating? Let's talk.
BuilderAI0 to 1ProductArchitectureGo-to-Market

Head of AI and Engineering

IOK Labs
2021 — 2025

NutriChef

Co-created from zero.
AI nutrition and coaching platform built on multi-agent LLM architecture.

  • Partnered with Ryan Fernando, India's leading sports nutritionist (Virat Kohli, Indian Cricket Team), to build and launch the RF Health app
  • Engineered multi-agent workflows for dietary profiling, meal planning, and blood report analysis
  • Shipped LLM-powered features — Snap-to-Track, Write-to-Track, Health Score, and more
  • Architected the core backend from scratch, cutting nutritionist workflow time by 10x
  • Delivered scalable multi-tenant infrastructure supporting multiple white-label clients simultaneously
LLMMulti-agent SystemsWhite-label SaaSBackend Architecture

tinychef (Sanjeev Kapoor Recipes)

Co-founded by Sanjeev Kapoor.
Scaled the Sanjeev Kapoor Recipes Alexa Skill and app to 100K+ MAU and 500K+ monthly voice interactions.

  • Partnered directly with the Amazon Alexa team as India's #1 recipe intent partner
  • Built AI-powered search and recommendation engine, lifting user engagement by 30%
  • Built in-house ads platform and CMS, turning brand partnerships into a standalone revenue channel
  • Worked closely with brands like ITC, Swastha Cookware, and Heritage on AI-powered campaigns and content
Amazon Alexa PartnerConversational AIAI Search & RecommendationsVoice Ads

Software Engineer, AI

Nayan Technologies
2020 — 2021

AI-powered traffic monitoring and road safety.
Deployed violation detection for Dubai Police through Dubai Future Accelerators.

  • Built traffic violation detection from CCTV footage, 91% accuracy in production across UAE
  • Delivered POCs for IdeaForge (drone-based parking violations) and Delhi Police (real-time number plate recognition against stolen vehicle databases)
  • Retrained YOLOv4 on custom datasets, improving accuracy by 25%. Reduced false violations by 40%
  • Built in-house data annotation engine, automating the full train-validate-test pipeline
Traffic & SurveillanceComputer VisionDeep LearningMLOps

AI Research Intern

Adventum
2020

Medical imaging AI. Piloted with ophthalmology clinics in Spain.
Built a pipeline to detect early-stage eye diseases from retinal scans.

  • Deployed working POC for real-time diagnostic applications with clinics
  • Developed AI models for detecting diabetic retinopathy and glaucoma from fundus and OCT scans
  • Applied transfer learning, ensemble methods, and multi-modal learning across various architectures to improve diagnostic accuracy
AI in HealthcareResearchOphthalmologyDeep LearningComputer Vision

Capabilities

Built at startups. Learned to do it all.

Working with startups from day one taught me to wear different hats.
Whatever the product needed to move forward, I picked it up and shipped it.

Full-Stack Builder

Started with AI research. Built the backend to serve the models. Learned cloud to deploy them. Picked up everything else along the way because that's what building a product takes.

Backend SystemsCloud & InfrastructureFrontend & WebAPI DesignSystem ArchitectureCI/CD

AI in Production

From detecting eye diseases in retinal scans to building multi-agent LLM systems that serve real users. Six years of shipping AI, not just training models.

Large Language ModelsMulti-agent WorkflowsComputer Vision & Deep LearningMLOps & Model Pipelines

Zero to One Product

No PM, no product team. Scoped features, ran sprints, prioritized roadmaps, and demoed to partners. Startups taught me product the hard way.

Product Roadmaps & PrioritizationScrum & Sprint PlanningUser Research & Feedback LoopsFeature Scoping & Demos

Growth and Revenue

Worked on brand partnerships. Captured partner requirements. Tracked revenue numbers. Built the tools to measure it all. At a startup, the business side is everyone's job.

Revenue Tracking & AnalyticsPartner Requirements & DeliveryBrand PartnershipsGo-to-Market Strategy

Activity

What I've been up to.

I ship code. Regularly.

What the last year looks like, commit by commit.

When I'm not building.

Trekking through mountains, chasing sunsets, or exploring new places.

Contact

Building something? Let's talk.

Happy to talk AI, tech, product, and everything in between. Grab a slot below or just drop a message.