Hiring Recommendation Engineers: The Complete Guide

# Recommendation Engineer

**Location:** San Francisco, CA (Hybrid) · **Employment Type:** Full-time · **Level:** Mid-Senior

## About [Company]

[Company] is a marketplace connecting buyers with unique products from 5 million independent sellers. With 100 million listings, personalization is critical—users need to discover the products they'll love without being overwhelmed by choice.

Our recommendation systems drive 40% of marketplace GMV. When recommendations work well, buyers find products they love, and sellers succeed. This is core infrastructure with direct business impact.

**Why join [Company]?**

- Work on recommendations at scale—100M listings, 50M monthly active users
- Direct business impact—recommendations drive 40% of GMV
- Solve hard ML problems (cold start, multi-objective, real-time)
- Join a 12-person RecSys team within a 300-person engineering org
- Series E company, profitable, industry-leading marketplace
- Competitive compensation with strong equity component

## The Role

We're looking for a Recommendation Engineer to help us build world-class personalization across our platform. You'll work on the full recommendation stack: candidate generation, ranking, real-time personalization, and experimentation.

This is classic RecSys engineering with real challenges: cold start for new users and items, multi-objective optimization (engagement, diversity, fairness to sellers), two-sided marketplace dynamics, and scale. You'll build systems that serve billions of recommendations daily.

The ideal candidate has experience building production recommendation systems and understands the unique challenges: collaborative filtering's sparsity problems, embedding model design, feature engineering, and multi-stage ranking. You're excited about experimentation and measuring what actually works.

## What You'll Work On

### First 30 Days
- Onboard to our recommendation stack, feature platform, and experimentation system
- Understand our recommendation surfaces (home, search, email, notifications)
- Ship your first recommendation improvement (new feature, model tweak)

### First 90 Days
- Own a recommendation surface or model component
- Run A/B experiments to validate improvements
- Contribute to our feature store and model serving infrastructure
- Collaborate with product on personalization roadmap

### First Year
- Lead recommendation initiatives with measurable business impact
- Architect improvements to our candidate generation or ranking systems
- Develop new recommendation surfaces or use cases
- Mentor teammates on RecSys best practices

## Objectives of This Role

- Improve recommendation quality metrics (CTR, conversion, engagement)
- Maintain diversity and discovery (not just popular items)
- Reduce cold start problems for new users and items
- Improve feature pipeline reliability and freshness
- Build experimentation practices that generate learnings

## Responsibilities

**Model Development (50%)**
- Build and improve recommendation models (collaborative filtering, embeddings, neural ranking)
- Design multi-stage retrieval and ranking pipelines
- Engineer features from user behavior, item attributes, and contextual signals
- Handle cold start problems for new users and items
- Optimize for multiple objectives (engagement, diversity, seller fairness)

**Infrastructure & Serving (30%)**
- Build candidate generation systems for retrieval at scale
- Implement real-time personalization with low latency requirements
- Contribute to feature store and model serving infrastructure
- Optimize recommendation pipelines for performance and cost
- Build monitoring for model quality and system health

**Experimentation & Analysis (20%)**
- Design and run A/B experiments for recommendation changes
- Analyze experiment results and translate to product insights
- Build offline evaluation frameworks for faster iteration
- Collaborate with product on recommendation strategy
- Document learnings and contribute to team knowledge

## Required Skills and Qualifications

**Technical Skills**
- 5+ years of ML engineering experience
- 3+ years building production recommendation systems
- Deep experience with collaborative filtering and embedding models
- Strong understanding of multi-stage retrieval and ranking
- Production ML system experience (feature stores, model serving)
- Strong A/B testing and experimentation skills
- Proficiency in Python and ML frameworks (TensorFlow, PyTorch)

**RecSys Knowledge**
- Understanding of cold start problems and solutions
- Experience with multi-objective optimization
- Familiarity with recommendation evaluation metrics

**Soft Skills**
- Strong analytical skills for experiment interpretation
- Collaboration with product and other ML teams
- Clear communication about complex technical concepts

## Preferred Skills and Qualifications

- E-commerce or marketplace recommendation experience
- Real-time ML serving experience (< 50ms latency)
- Experience with large-scale feature stores (Feast, Tecton)
- Two-sided marketplace recommendation experience
- Experience with graph-based recommendations
- Publications at RecSys, KDD, or similar venues

## Tech Stack

**ML:** Python, PyTorch, TensorFlow, XGBoost for ranking

**Embeddings:** Learned embeddings, FAISS for ANN search, Pinecone

**Feature Store:** Feast, custom real-time features

**Infrastructure:** Kubeflow for training, custom serving, Kubernetes

**Data:** Spark, Snowflake, Airflow for batch pipelines

**Experimentation:** Custom A/B platform, Databricks for analysis

## Compensation and Benefits

**Salary:** $170,000 - $230,000 (based on experience)

**Equity:** 0.05% - 0.15% (4-year vest, 1-year cliff)

**Benefits:**

*Health & Wellness*
- Medical, dental, and vision insurance (100% covered)
- $100/month wellness stipend
- Mental health support through Lyra

*Time Off*
- Unlimited PTO with 15-day minimum encouraged
- 11 paid company holidays
- 16 weeks paid parental leave

*Professional Development*
- $3,000 annual learning budget
- Conference attendance (RecSys, KDD, SIGIR)
- Internal ML reading groups and tech talks

*Financial*
- 401(k) with 4% company match
- Equity refresh grants annually

*Workspace*
- $1,500 home office setup allowance
- Hybrid work (2-3 days in SF office)
- MacBook Pro and GPU options for local development

## Engineering Culture

**How we work:**
- Experiment-driven development—measure everything
- Full-stack ownership of recommendation surfaces
- Weekly RecSys team syncs to share learnings
- Code review and model review for all changes
- Blameless post-mortems after issues

**What we value:**
- User experience over engagement metrics
- Diversity and discovery, not just popular items
- Fairness to sellers in marketplace recommendations
- Scientific rigor with business pragmatism
- Learning from experiments, even negative results

## Interview Process

Our interview process typically takes 2 weeks.

- **Step 1: Recruiter Screen** (30 min)
  Background discussion and role overview.

- **Step 2: Technical Screen** (60 min)
  RecSys fundamentals and ML concepts.

- **Step 3: RecSys Deep Dive** (90 min)
  Technical discussion about recommendation system design, your experience, and how you approach RecSys challenges.

- **Step 4: ML Coding** (90 min)
  Practical ML exercise—feature engineering, model implementation, or evaluation.

- **Step 5: System Design** (60 min)
  Design a recommendation system: retrieval, ranking, serving, experimentation.

- **Step 6: Team Interviews** (2 x 30 min)
  Meet potential teammates. Discussion about collaboration and working style.

---

*[Company] is an equal opportunity employer.*