Research Roadmap for Meridian DSP's Bid-Matching Pipeline
Overview
What this challenge is about.
Assemble a 30-40 paper reading list on retrieval, ranking, and fairness for a DSP pipeline, then propose two research investments. Earn a verifiable certificate.
The scenario
Meridian DSP competes in real-time advertising auctions where its bid-request matching pipeline must score and rank candidate ads within a tight per-request latency budget across billions of daily requests. The ML team has limited research capacity and can fund only two exploratory investments next year, so the choice must be evidence-led.
The Brief
What you'll do, and what you'll demonstrate.
Decide and justify which two research directions Meridian DSP should invest in next year to improve its bid-request matching pipeline, grounded in a synthesis of recent information-retrieval and ranking research.
Earning criteria — what you'll demonstrate
- Synthesize a body of recent research into themes and a defensible point of view rather than paper-by-paper summaries.
- Distinguish techniques that are production-ready from those that are still speculative for a real ranking system.
- Tie research findings to the concrete latency, scale, and data constraints of a specific pipeline.
- Communicate a prioritized technical agenda persuasively to a non-uniform stakeholder audience.
- Reason about ranking bias and fairness as first-class engineering and business concerns.
Program Fit
Where this fits in your program.
Sharpens the same skills your degree expects you to demonstrate.
Aligned coursework coming soon.
Skills
Skills you'll demonstrate.
Each one shows up on your verified credential.
- Information Retrieval
Apply information retrieval to solve real industry problems and demonstrate production-level capability.
- Learning To Rank
Apply learning to rank to solve real industry problems and demonstrate production-level capability.
- Research Synthesis
Apply research synthesis to solve real industry problems and demonstrate production-level capability.
- Dense Retrieval
Apply dense retrieval to solve real industry problems and demonstrate production-level capability.
- Stakeholder Communication
Apply stakeholder communication to solve real industry problems and demonstrate production-level capability.
- Ranking Bias
Apply ranking bias to solve real industry problems and demonstrate production-level capability.
Careers
Career paths this challenge builds toward
Completing this challenge demonstrates skills that transfer directly to these roles:
Machine Learning Engineer, Ranking
Translating current retrieval and ranking research into shippable pipeline changes is the daily work of a ranking ML engineer; this challenge builds the muscle of judging what is production-ready and tying it to system constraints.
This challenge sharpens
- information-retrieval
- learning-to-rank
- dense-retrieval
Applied Research Scientist
Applied researchers must survey a fast-moving field, form a defensible point of view, and propose where a team should invest; this challenge mirrors that scoping-and-synthesis workflow end to end.
This challenge sharpens
- research-synthesis
- information-retrieval
- ranking-bias
Technical Lead, Search and Recommendations
Leads prioritize limited research capacity and sell those bets to stakeholders; this challenge practices turning evidence into a prioritized roadmap and a persuasive readout for a working team.
This challenge sharpens
- learning-to-rank
- research-synthesis
- stakeholder-communication