Staff ML Engineer at Espresso AI (New York, NY)

Espresso AI· New York, NY· $225k–$450k·6d ago

Role details

Salary

$225k–$450k

Work type

Onsite

Employment

Full-Time

Equity

Yes

Job description

Train cutting-edge models to optimize our users' compute costs.

High Hiring Bar & Elite Peers: The company explicitly targets “the strongest group of engineers you’ve ever worked with.” Staff engineers from Google, Apple, and MIT form the core team.
Early-Stage Impact: With ~15 people and strong product-market fit, new hires get substantial equity and can significantly shape the company’s trajectory. New engineers are expected to ship a meaningful feature in their first week.
Compensation: Competitive salary + meaningful equity. Engineering roles: $225,000–$450,000 salary + equity. Sales (OTE): $200,000–$350,000.
Office Culture: 5 days/week in-office in Brooklyn, NYC. Relocation support is available.
Autonomy & Ownership: The company values engineers who work autonomously, prioritize ruthlessly, and communicate effectively. Sales hires are expected to help build the sales process from the ground up.
Cutting-Edge Work: Engineers train cutting-edge models that answer questions like “how fast will this code run?” and “can this SQL query run faster on larger infra?” — blending AI, distributed systems, and compiler-style optimization.

One-liner: Espresso AI uses generative AI and neural optimization to automatically reduce cloud data warehouse costs for Snowflake and Databricks, saving customers up to 70% on their compute bills.
Entity Type: Private (Seed stage)
Headquarters: Brooklyn, New York, USA
Founded: 2023
Founders: Ben Lerner (CEO & Co-Founder), Alex Kouzemtchenko (CTO & Co-Founder), Juri Ganitkevitch (Co-Founder)

Primary Industry: Cloud cost optimization / AI-powered data infrastructure
Target Customers: B2B — enterprises and SMBs using Snowflake and Databricks who face rising compute costs
Mission: To make AI accelerate compute by 1000x, starting with data warehousing, by building the world’s first neural optimizer.

Espresso AI Neural Optimizer (Core Product): An automated, ML-driven platform that optimizes Snowflake and Databricks data warehouses in real-time. The system runs autonomously, applying performance engineering and scheduling optimizations to cut compute costs by up to 70%. Set-up requires a single SQL command.
Kubernetes for Snowflake: A product launched in 2025 that uses Kubernetes to renovate and optimize Snowflake data warehouses, further reducing costs and improving performance.

Valuation/Market Cap: Not publicly available
Key Metric: Total funding of $11.0M (Seed round, led by Nat Friedman and Daniel Gross, announced May 2024)
Notable Investors/Partners: Nat Friedman, Daniel Gross, FirstMark Capital (Matt Turck), and angel investors including Tristan Handy (dbt Labs founder)
Growth Signals:
- 15 employees (+35.7% YoY)
- Multiple AEs actively closing deals — the company states they have product-market fit
- Expanding support from Snowflake-only to include Databricks
- Strong inbound interest and active hiring for engineering and sales roles

First Neural Optimizer: Espresso AI is building what it calls the world’s first neural optimizer — using LLMs to deeply understand and optimize SQL code, scheduling, and infrastructure, leapfrogging traditional rule-based optimization tools.
Elite Team: Founders and engineers come from Google (NLU for Search), Apple (Siri), and MIT, with deep expertise in performance engineering and AI.
Guaranteed ROI Model: Only charges customers based on savings — no onboarding costs, no minimums, no commitments — creating strong alignment and trust.
Autonomous 24/7 Operation: Unlike hiring a team of data engineers, the platform runs continuously, requiring minimal human intervention after setup.

Near-term: Deepen Snowflake and Databricks optimization capabilities; scale sales team to capture growing demand; hire staff-level engineers to build out distributed systems and ML infrastructure.
Long-term: Expand beyond data warehousing to optimize all major sources of cloud compute, making AI-driven compute optimization a universal infrastructure layer.

Sources