In today’s digital learning ecosystem, discoverability is everything. Learners are flooded with content — but only find value when the right learning surfaces at the right moment. That’s where search relevance becomes critical.
Why Elasticsearch Matters in Enterprise Learning
Elasticsearch is a fast, scalable, and widely adopted search engine that sits behind many experiences we rely on daily — product filters on e‑commerce sites, autosuggest in apps, personalized recommendations, and even log dashboards used by engineering teams.
For enterprise learning platforms like Infosys Wingspan, Elasticsearch plays an equally important role. It helps learners navigate vast libraries containing hundreds of thousands of courses, videos, labs, and certifications. But while the basic information retrieval solution provides speed and scale, enterprises need more than fast results — they need meaningful, context-aware, and highly relevant results.
Because when learners can’t quickly locate what they need, engagement drops, outcomes diminish, and the learning experience loses its impact.
The Challenge: Delivering Relevant Learning Content at Scale
Traditional keyword search often retrieves information that is technically correct — but not truly helpful. Learners differ by role, proficiency, location, preferred language, and their unique learning journeys. At this scale, relevance cannot be left to chance.
We needed a search system that understands:
- Intent — what the learner is truly looking for
- Behavior — what they tend to engage with
- Context — who they are and where they are in their learning journey
This led us to reimagine search ranking through a multi-layered, intelligence-driven approach.
Our Solution: Multi‑Layer Reranking
We built a 5‑step reranking pipeline that blends traditional information retrieval, machine learning, personalization, and business logic into a unified ranking strategy.
1. BM25 Retrieval
Quickly fetches an initial candidate set of relevant results.
2. Learning to Rank (LTR) with XGBoost
Trained using:
- Rich telemetry (clicks, progress, completion)
- A Dynamic Bayesian Network (DBN) click model
- Metadata and user‑behavior signals
This allows the algorithm to understand what learners actually find useful.
3. Trending Content Boosts
Highlights what’s gaining traction across the learner base.
4. Personalization
Adjusts ranking based on learner‑specific attributes — language, location, history, and interests.
5. Business Priority Boosts
Elevates strategic programs, certifications, or initiatives when needed.
Each layer incrementally improves precision, ensuring only the most relevant results rise to the top.
The Data Behind the Model
To train the system effectively, we used:
- 174,000+ query–document pairs
- Four‑level graded relevance
- Extensive feature engineering across content metadata and behavioral insights
This comprehensive dataset helped our LTR model gain a deep understanding of learner preferences and content utility.
The Outcomes: Clear, Measurable Impact
Deploying our multi‑layer reranking architecture produced significant improvements in learner experience and platform efficiency:
- NDCG@10: improved from 0.913 → 0.956
- MRR@10: more than 2× increase
- Average click position: improved from 5.3 → 4.7
- More clicks shifted to top results
- Overall searches increased by 44% over six months
In essence, learners found relevant content faster, and trust in the search experience increased substantially.
What’s Next?
We’re already exploring the next frontier of intelligent learning search:
- Hybrid retrieval combining BM25 + dense vector embeddings
- Reinforcement learning that adapts ranking from ongoing learner feedback
- Deeper, real‑time personalization
- Unified search relevance pipelines across all our learning platforms
Our vision is simple:
A search system that continuously learns and evolves—just like our learners.
How This Strengthens Infosys Wingspan
At its core, Infosys Wingspan is designed to deliver personalized, scalable, enterprise‑ready learning experiences. Intelligent search is a cornerstone of this vision.
By integrating multi-layer reranking and continuous learning signals into Wingspan:
- Learners discover relevant content quickly
- Organizations drive higher adoption, completion, and skill mobility
- Learning journeys become intuitive, frictionless, and meaningful
Smarter search isn’t just a feature—it’s a competitive advantage.
If you’re looking to transform how your workforce learns, discovers, and grows, Infosys Wingspan is built to propel your enterprise into the future of AI-powered learning.
Dive Deeper
Peer‑Reviewed Research: https://dl.acm.org/doi/10.1145/3627217.3627224
Talk Recording: https://www.elastic.co/events/elasticon/archive/elasticon-bengaluru/improving-search-relevance-using-multilayer-reranking
