Member of Technical Staff (Machine Learning Engineer, Ranking Quality

About the role

Perplexity is seeking an experienced Machine Learning Engineer to elevate search quality across the middle and later stages of ranking. The role targets a ranking generalist capable of owning ambiguous problems from initial definition through final deployment. Exceptional depth in either neural ranking or production ranking systems is essential for success in this position. You will define ambiguous search challenges and own them from first metric to deployed safeguard while navigating complex trade-offs. Experiments will target middle and later ranking stages while protecting real user pathways and system integrity. This role suits a ranking generalist who moves models, data, and process in one coherent arc from conception to production. You will act as the single point of ownership for ranking quality, ensuring alignment between technical execution and user experience.

Key facts

What you'll do

Design intake criteria for ranking experiments and align metrics with unclear product questions to establish clear evaluation frameworks. Construct neural and classic ranking components while tracing behavior through live traffic corridors to validate real-world performance. Shape retrieval signals and classification logic inside multi-stage setups that respect strict latency boundaries and operational constraints. Govern feature pipelines and inference paths so ranking cascades stay observable, maintainable, and low in delay across all stages. Balance accuracy, response time, stability, and implementation cost in every architecture choice to optimize total system value. Coordinate with data, platform, and product groups while signing off on final ranking outcomes to ensure accountability and quality ownership. Maintain monitoring views and evaluation harnesses that expose regressions before users notice, enabling proactive system improvements. Partner with infrastructure teams to harden deployment and rollback for ranking-critical services to ensure resilience and reliability. Continuously iterate on evaluation methodologies and data pipelines to drive incremental improvements in search quality over time.

Requirements

You bring deep evaluation methods for search and recommender pipelines into daily work, applying rigorous experimental design and analysis. Experience owning large-scale production ranking stacks or broad quality problem sets must be evident through concrete examples and outcomes. Data, model, serving, and monitoring skills should move complex projects forward without constant breakdown or dependency escalation. You drive vague cross-team initiatives while skipping endless task breakdown requests, demonstrating independence and ownership. Neural ranking approaches or low-latency runtime expertise has to shape major design decisions that impact system architecture and user experience. At least five years of relevant industry practice is necessary to bring context, judgment, and maturity to complex challenges. You possess a deep understanding of search or recommender systems and their evaluation methodologies, including offline and online testing frameworks. Demonstrated proven ownership of ambiguous, cross-team initiatives without requiring continuous task decomposition is essential for this role. Maintain exceptional depth in either modern neural ranking methods or low-latency ranking systems and runtime to guide technical trade-offs. Show strong machine-learning and software-engineering skills across data, models, serving, and monitoring to build robust, scalable solutions.

Nice to have

No preferred skills are specified beyond the core requirements and qualifications listed in the original source material. The role description contains no additional Nice to have section, and all preferences are implicitly covered by the stated requirements and qualifications. Any skills beyond those documented would be inventions not supported by the source.

Practical notes

The position is based in Belgrade and requires full-time engagement. The compensation package is specified as 140 000 to 200 000 USD per year for this role. There are no additional details regarding working hours, travel expectations, visa sponsorship, or application deadlines provided in the source material. All practical information is limited to what is explicitly stated in the original job description from the source.