Machine Learning Engineer

About the Role

As a Machine Learning Engineer, you will design and develop the ML systems fueling Firecrawl, one of the fastest-growing AI data products. Your core focus will be improving search through advanced ranking and relevance models, starting from feature engineering and model training to deploying and maintaining these models in production. Beyond search, this role extends ML capabilities to extraction quality, content classification, and LLM-based product features.

You will also lead the measurement strategy by creating A/B testing frameworks and evaluation metrics that guide product launch decisions and continuous improvement.

Company Overview

Firecrawl transforms the web into accessible data via a single API call converting URLs into structured, LLM-ready formats. The product has experienced explosive growth, achieving eight figures in annual recurring revenue within the first year and doubling in the second, backed by a top 50 GitHub repository with over 180,000 stars.

Operating from San Francisco and a new Toronto hub, the team is small, highly skilled, and empowered to take full ownership of impactful projects in a fast-moving AI infrastructure environment.

Key Responsibilities

• Enhance ranking and relevance for Firecrawl Search through end-to-end ML development.
• Develop and optimize learning-to-rank, query understanding, and LLM-based retrieval models.
• Apply ML techniques across products for extraction accuracy, classification, and LLM feature evaluation.
• Analyze large-scale query logs and user behavior data to identify product strengths and weaknesses.
• Construct scalable data pipelines converting vast web crawl and query data into training features.
• Collaborate with platform, search, and cloud DevOps teams to deploy models efficiently and cost-effectively in production.
• Design and implement A/B testing frameworks and offline evaluation methods for measuring product impact.
• Partner on product rollouts by defining success metrics, analyzing experiment results, and making informed launch decisions.
• Monitor post-launch performance and iterate based on empirical findings.

Required Qualifications

• Minimum 3 years of experience delivering ML models into production environments and managing their lifecycle including deployment, monitoring, and retraining.
• Proven expertise in ranking or relevance modeling such as learning-to-rank, recommendation systems, or search quality improvements.
• Comfortable handling and processing large-scale, data-intensive systems and logs that exceed in-memory capacities.
• Proficient in writing robust production-grade code, particularly in Python, and working within complex backend infrastructures.
• Experienced in rigorous experimental design including A/B testing and statistical analysis to assess genuine performance lifts.
• Effective communication skills to clearly report results and recommendations to cross-functional teams.

Preferred Skills

• Background in MLOps tools including MLflow, experiment tracking, model registries, or feature stores; Kubernetes knowledge is advantageous.
• Experience establishing or refining experimentation platforms in previous roles.
• Familiarity with embedding models, vector search, and LLM-based relevance evaluation approaches.
• Skilled in evaluating large-scale LLM outputs such as quality scoring, structured data extraction, and agent behaviors.
• Experience with big data processing frameworks like Apache Spark.

What This Role Is Not

• This is not suited for individuals focused solely on statistical analysis or data science without production deployment and engineering focus.
• Not intended for those seeking narrow specialization or delegating responsibility for model deployment and maintenance.
• Unsuitable for candidates prioritizing process over rapid and impactful delivery.

Work Environment and Culture

The company operates at an intense pace motivated by the urgent market opportunity in AI infrastructure. Employees are expected to embrace ownership, fast iteration, and deep technical craftsmanship.

Compensation and Benefits

Compensation is competitive and impact-driven, with differing salary ranges depending on location:

• San Francisco: $250,000 to $290,000 USD annually
• Toronto: $210,000 to $224,000 CAD annually

Equity stakes are granted competitively.

Employee benefits include:

• Mandatory minimum of 15 PTO days with flexibility for additional leave
• Parental leave offering 12 weeks of full pay for all parents
• Monthly $100 wellness stipend for fitness or mental health
• $1,000 annual learning and development budget
• Team offsites and a sabbatical after four years (3 months paid)

Healthcare and related benefits vary by location:

• US-based employees receive comprehensive medical, dental, vision coverage, employer-paid life & disability insurance, virtual and mental health services, fertility support, employee assistance programs, retirement plans (401(k)), and pre-tax benefits.
• Canadian employees have extended health, dental, vision benefits through Manulife, virtual care, mental health support, fertility programs, and group RRSP retirement plans.

Office perks:

• San Francisco HQ offers snacks, drinks, team lunches, ping pong, and startup energy, plus access to an E-Bike loan program.
• Toronto Hub provides snacks, drinks, team lunches, scenic views, commuter transit cards (PRESTO), and parking benefits.

Hiring Process

• Submit work samples demonstrating delivered models, search systems, experimentation frameworks, or data pipelines along with motivation for the role.
• Introductory conversation (~25 minutes) covering background, interests, and candidate questions.
• Technical discussion (~45 minutes) focused on tackling real Firecrawl challenges and problem-solving approaches.
• Founder meeting (~25 minutes) to align on culture, working style, and pacing.
• Paid trial project (~40 hours) involving hands-on collaboration on a product task, providing realistic insight into team workflow and delivery expectations.
• Rapid decision and offer following trial completion.