As a four-time founder, angel investor, and former VC, Yanda Erlich brings a unique perspective to building Weights & Biases, a leading MLOps platform. He advises founders to build painkillers, not vitamins, and to rigorously evaluate their own companies from an investor's perspective. For making tough choices, he recommends a 'regret minimization' framework to focus on long-term outcomes.
Key takeaways
- Build developer tools that solve a painful, expensive problem.
- Before you pitch VCs, audit your company like an investor would.
- Frame your traction in terms of velocity and capital efficiency.
- For hard decisions, ask which outcome you will regret more in five years.
- A great founding team is non-negotiable for top-tier investors.
- Don't just build a product; build a system for distribution.
Yan-David (Yanda) Erlich has co-founded four venture-backed startups, worked as an angel investor, and served as a partner at a venture capital firm. Now he's the co-founder and CEO of Weights & Biases, an MLOps platform valued at over $1 billion with backing from NVIDIA, Insight Partners, Felicis Ventures, and Coatue.
Drawing from his experience on both sides of the table, this is a tactical guide for founders based on the lessons from his journey. We'll cover what top investors look for, the non-obvious traits of a billion-dollar opportunity, and a framework for making the hard decisions you'll face while building your own company.
What Investors Look For: A Founder's Checklist
Most pitch advice is about storytelling. That's important, but top investors apply a systematic financial and strategic lens to every deal. Yanda’s experience as a VC reveals what happens after you leave the room. Before you even start fundraising, you should audit your own startup with the same rigor.
1. Is it a Painkiller or a Vitamin?
Investors want to back painkillers. Vitamins are 'nice to have'—they offer incremental improvements. Painkillers solve an urgent, expensive, and unavoidable problem.
Vitamin: A tool that makes developers slightly happier or a workflow 10% more efficient. · Painkiller: A tool that unblocks a critical business function, prevents catastrophic failures, or is essential for shipping the company's core product.
Weights & Biases is a painkiller for companies building with AI. Without a system to track experiments, version models, and ensure reproducibility, ML development becomes a chaotic, expensive mess. It moves from a 'nice to have' to a 'must have' as soon as a company gets serious about AI. Ask yourself: is your product a convenience or a necessity?
2. The Market: Is the Tide Lifting Your Boat?
Even a great product in a small or shrinking market will struggle. Investors look for founders building in markets with powerful tailwinds. For W&B, that tailwind is the explosion of AI development.
TAM (Total Addressable Market): How big is the total potential market? (e.g., 'All companies spending on cloud infrastructure') · SAM (Serviceable Addressable Market): Which part of that market can you realistically reach? (e.g., 'Companies with dedicated machine learning teams') · SOM (Serviceable Obtainable Market): What can you capture in the next 3-5 years? This is your business plan.
Top VCs want to see a path to building a $100M+ revenue business, which typically requires a multi-billion dollar TAM. For a dev tool, you can estimate this by multiplying the number of potential developer users by a realistic annual contract value (ACV).
3. Traction: The Language of Investor Trust
Traction is evidence you're on the right track. It's the most powerful way to de-risk the investment for a VC. What counts as 'good' traction changes by stage:
Pre-Seed: A credible technical founder with a prototype and deep knowledge of the problem. Early design partners or a waitlist can be a strong signal. · Seed: The first $100k-$500k in Annual Recurring Revenue (ARR). The key is showing velocity —how quickly you acquired those first customers. Going from $0 to $20k MRR in 4 months is more impressive than taking 18 months. · Series A: Typically $1M to $3M in ARR. Investors are looking for a repeatable GTM motion. Can you prove that if you put $1 into sales and marketing, you get more than $1 back in predictable revenue?
Founder Mistake: Many founders show a cumulative revenue chart that only goes up and to the right. Investors see right through this. They want to see monthly or quarterly new revenue, churn rates, and growth persistence. Be transparent about the underlying metrics.
The Founder's Mindset: Optimize for Regret
Building a startup is a series of high-stakes decisions under uncertainty. Yanda speaks about a powerful mental model for navigating these choices: the Regret Minimization Framework.
The framework is simple. When facing a difficult choice, project yourself five or ten years into the future and look back on the decision you're about to make. Then ask: 'Which course of action will I regret the least?' This isn't about avoiding failure. It's about ensuring you take the path that aligns with your long-term convictions, even if it's harder in the short term.
When to Use This Framework
Pivoting vs. Persevering: Your current product isn't working. You have a new, exciting idea, but it means abandoning months of work. Ask: In five years, will you regret not trying the new idea more than you'll regret giving up on the original one? · Making a Key Hire: You have a candidate who is perfect on paper but a potential cultural mismatch. Hiring them could accelerate your roadmap, but passing means a long, painful search. Ask: In five years, will you regret a toxic culture that drove away your best people more than you'll regret a 3-month product delay? · Fundraising Decisions: You have a term sheet from a Tier 2 fund, but you think you might have a shot with a top-tier firm if you wait. Waiting risks running out of money. Ask: In five years, will you regret having the 'wrong' partners for the life of your company more than you'll regret the stress of a few extra weeks of fundraising?
This framework forces you to separate temporary pain from long-term misalignment. It pulls you out of the day-to-day panic and connects you back to the fundamental reason you started the company.
How to Apply This This Week
You don't need to be raising a mega-round to benefit from these lessons. Here are three actionable steps you can take right now.
Run an Investor Audit on Your Own Company. Create a private document and honestly assess your startup against the criteria: market size, traction velocity, and product differentiation (painkiller vs. vitamin). Where are you weakest? What one metric could you improve in the next 30 days to make the story stronger? · Define the 'Painkiller' ROI. Write down a simple, three-bullet-point explanation of how your product solves an urgent and expensive problem for your customers. Quantify it if you can (e.g., 'Saves 10 engineering hours per week,' 'Reduces server costs by 30%'). This is the core of your pitch to customers and investors. · Identify Your Next 'Regret Minimization' Decision. What's the biggest, most stressful decision on your plate right now? Instead of thinking about the immediate pros and cons, write down your answer to the question: 'In 2029, which outcome will I wish I had chosen?' Use that as your guide.
Frequently asked questions
- What is MLOps and why is it important?
- MLOps (Machine Learning Operations) is the practice of standardizing and streamlining the machine learning lifecycle. It's critical because building and deploying reliable AI models at scale is complex, and MLOps tools provide the infrastructure for tracking experiments, managing data, and monitoring models.
- What key metrics do investors look for in a Series A software startup?
- At Series A, investors typically want to see $1M-$3M in Annual Recurring Revenue (ARR), strong month-over-month growth (15-20%+), high gross margins (75%+), and early signs of product-market fit like low churn and a repeatable sales process.
- What is the 'regret minimization framework' for founders?
- It's a mental model for making major decisions (e.g., pivoting, hiring a key exec, fundraising). You project yourself 5-10 years into the future and ask: 'Which path, if it fails, will I regret the least?' It helps prioritize long-term vision over short-term pain or fear.
- What's a common mistake when pitching developer tools to VCs?
- A common mistake is focusing exclusively on the technical features instead of the business impact. Founders must translate elegant code or a powerful platform into clear ROI for the customer: saved engineering hours, faster time-to-market, or enabling new product lines. Investors fund businesses, not just technologies.
