Senior Machine Learning Engineer, Ads Response Prediction
About this role
The Ads Response Prediction team at Instacart owns all systems, algorithms, and machine learning models required to deliver a relevant and engaging ads experience across the platform. This group focuses on search and exploration retrieval systems, sequential modeling for next-interaction recommendations, relevance models, pCTR models, and bidding frameworks to balance customer satisfaction, advertiser outcomes, and revenue.
As a Senior Machine Learning Engineer on this team, the focus centers on research and development for pCTR and conversion prediction models, theoretical problem formulation, training methodology, and model quality. Responsibilities include mitigating data biases such as selection and position bias, advancing multi-task learning and sequence modeling capabilities, and contributing to foundation model approaches like generative retrieval systems and semantic IDs. Strong candidates bring a background in applying machine learning to ranking, recommendation, or prediction at scale, along with a talent for translating ambiguous problems into clear research directions.
What you will do
- Own and execute the research and development of pCTR and conversion prediction models to improve calibration and accuracy across ads surfaces.
- Design and implement debiasing techniques such as Mixed Negative Sampling, Inverse Propensity Weighting, and counterfactual risk minimization.
- Contribute to next-generation Multi-Domain Multi-Task model architectures incorporating Mixture-of-Experts and Transformer layers for sequential user behavior.
- Work on sequence modeling initiatives, including the TIGER generative retrieval system and Semantic ID representation learning across multiple placements.
- Formulate and scope ambiguous modeling problems from first principles and translate business observations into well-defined machine learning research directions.
- Collaborate with the broader machine learning community on foundation models using autoregressive user behavior prediction.
- Contribute to a culture of technical rigor through design reviews, paper sharing, and experiment retrospectives.
What you need
- Master's or PhD in machine learning, statistics, computer science, information retrieval, or a closely related quantitative field, or equivalent experience.
- Minimum of 3 years of combined academic and industry experience applying machine learning to ranking, recommendation, or prediction problems at scale.
- Deep understanding of CTR and conversion prediction modeling, including familiarity with architectures such as Deep & Wide, DeepFM, DCN, and multi-task learning formulations.
- Strong foundation in causal inference, counterfactual reasoning, and training data bias mitigation.
- Proficiency in Python and deep learning frameworks like PyTorch, TensorFlow, or JAX, alongside fluency in data manipulation tools such as SQL, Spark, and Pandas.
- Demonstrated track record of formulating ambiguous problems into well-scoped research directions and delivering results through rigorous experimentation.
- Strong written and verbal communication skills with the ability to explain complex modeling decisions to cross-functional stakeholders.
Nice to have
- Experience in ads ranking or auction-based systems including pCTR, bid optimization, and ROAS feedback loops.
- Hands-on experience with autoregressive sequence models for user behavior prediction, generative retrieval, or transformer-based ranking architectures.
- Familiarity with learned representations such as Semantic IDs or product embeddings to reduce feature cardinality and cold-start challenges.
- Experience with transfer learning or domain adaptation techniques applied to recommendation or ranking models.
- Publication record in top-tier venues such as KDD, WWW, RecSys, NeurIPS, ICML, or SIGIR.
Benefits
- Market-competitive compensation
- New hire equity grant as well as annual refresh grants
- Flex First remote work policy with location flexibility
Skills and tech stack
Questions candidates ask
What is the remote work policy for this role?
Instacart operates under a Flex First remote work policy where employees have the flexibility to choose where they do their best work while staying connected through regular in-person events. Currently, the company is only hiring in the Canadian provinces of Ontario, Alberta, British Columbia, and Nova Scotia.
What is the salary range for this position?
The base pay range for a successful candidate based in Canada is $180,000 to $190,000 CAD. Offers may vary based on factors such as candidate experience and skills required for the role.
What kind of infrastructure and support does the team provide?
The team has strong machine learning infrastructure and MLOps support, including Delta/DBT-Spark data pipelines, Ray-based distributed training, and automated model deployment. This allows engineers to focus their energy on advancing modeling science rather than building infrastructure.
What does the compensation package include beyond base salary?
Additionally, this role is eligible for a new hire equity grant as well as annual refresh grants. The company also provides highly market-competitive compensation and benefits in each location where employees work.