Expert Apache Solr Consulting
Regardless of the size of your organisation, Sease’s team of professionals has what it takes to make your search project a success. Equipped with extensive technical experience and deep knowledge of the internals of the tech, our Apache Solr specialists guarantee excellent results at every stage of the design and development of your Apache Solr-based search system and in the integration of new techniques related to artificial intelligence.
Take a look at our Apache Solr consultancy services.
Why choose Sease as your Apache Solr experts?
Choosing us as your Apache Solr experts ensures expertise and dedicated support for optimizing your search capabilities.
With +10 years of expertise and a proven track record of successful implementations across diverse industries, our team brings in-depth knowledge and experience to every project.
From comprehensive consulting and architecture design to seamless integration and ongoing support, we provide end-to-end solutions that drive tangible results for your organization.
Alessandro Benedetti
DIRECTOR @ SEASE
Alessandro Benedetti is an Apache Lucene / Apache Solr Committer, and an Apache Solr PMC Member (Project Management Committee).
choose our services as the world’s most innovative companies do:
Apache Solr Analysis & Audit
Our Apache Solr exploration and discovery service provides tailored insights and strategic recommendations to optimize your search platform.
We offer expert advice on the architecture, configuration, and usage to enhance user experience and accommodate future needs.
You will get a detailed report, organised to highlight what you have done well and what could have been done better, in a readable and comprehensive format, including configuration/queries examples and corrections.
And the cherry on the cake, we’ll follow up with a call to clarify the report and respond to questions.
Apache Solr Migration Services
migration to apache solr
- Evaluation of pros and cons of migrating from a different search technology to Apache Solr.
- Migration from Elasticsearch, OpenSearch, Vespa and other search technologies to Apache Solr.
apache solr version upgrade
- Evaluation of pros and cons of upgrading to another version of Solr.
- Upgrade your current Solr version to the latest one
Apache Solr systems design & development
design & development
We undertake the technical design and development of complete search systems or specific features, tailoring solutions to meet our client’s project requirements. We aim to deliver efficient and reliable solutions that exceed expectations.
We are domain agnostic: from an enterprise search engine to e-commerce, from Music to TV shows and economics: we have a deep theoretical and practical understanding of information retrieval techniques and have successfully applied them to Solr in many different domains over the years.
Artificial Intelligence integration to Apache Solr
large language models
Large Language Models can be a powerful allied to power your Solr search.
Our information retrieval experts can guide you through selecting the optimal Large Language Model (LLM) for your project and domain, customizing it to suit your specific needs, and quantifying the enhancements achieved.
- Identify the best Large Language Model for your project
- Customize the LLM and integrate it into your search system
- Evaluate the enhancements achieved with the integration of the Large Language Model.
We have been exploring various LLM integrations such as:
- Natural Language to structured queries (see our talk at Berlin Buzzwords and SMDX)
- Apache Solr Neural Highlighting plugin
There are many ways in which our team of Apache Solr experts can enhance your Solr experience using Large Language Models.
Contact us to find out how to proceed with your project and domain.
retrieval augmented generation
Retrieval Augmented Generation opens the door to a completely new user experience where the top-k results of your Solr search are used to power a generative step that provides answers with citations in natural language.
Our team specialize in implementing and evaluating RAG, relying on Apache Solr as your main search backend.
- We evaluate the pros and cons of the implementation of a Retrieval Augmented Generation system into your project
- We integrate and optimize the RAG system into your search engine
neural (vector-based) search
We contributed vector-based search to Apache Solr!
Since then we’ve been extending and maintaining the official feature introducing improvements based on the experience we are accumulating with our clients.
From chunking to the choice of the best embedding model for your use case, our Apache Solr expert can help you implement vector-based search with Apache Solr.
HYBRID SEARCH
Hybrid search techniques allow you to mix traditional keyword search with vector-based search and get the benefits of both worlds.
We are contributing new functionalities in Solr about this (see our talks at Berlin Buzzwords and Community Over Code EU), and our team of Apache Solr experts can help you implement it in your use case.
Learning to Rank
Learning To Rank adds the ability of integrating Machine Learning models to learn better ways of ranking your search results based on many relevance, users’ and business factors.
We’ve been working hands-on in the Learning To Rank Solr module since 2015 and we were among the founder group (with Bloomberg).
Since then we’ve been actively maintaining and extending the functionality and presenting at conferences many times in this regard (see our blog posts about learning to rank).
We applied Learning To rank in many clients’ projects and we would be delighted to help you in implementing it in your Apache Solr instance.
natural language processing
We’ve been integrating NLP techniques in Solr since 2013, with a focus on Named Entity Recognition and Linking, Query intent classification, etc.
With the advent of modern models such as GliNER, this is still incredibly relevant.
We can help you designing and implementing such integration with your Solr instance.
query/document expansion
Many techniques can implemented in Apache Solr to enrich and expand your query and your document for a better recall and user experience.
Our Apache Solr experts can help you integrate learned sparse vector models into your Solr search and get the benefits and explainability of such interesting approaches!
multi-modal search
Multi-modal search allows you to search across different media such as text, video, audio and images.
We can help you introduce this functionality in your Solr search engine or improve your current solution.
Search Relevancy Tuning in Apache Solr
SEARCH RELEVANCE
Returning relevant search results is hard, but we are up for the challenge!
We’ve been working on improving the relevance of search systems for many years and in many different domains, we explored many techniques, presented approaches at conferences and hosted many pieces of training on the topic.
We can tune the search relevancy of your Apache Solr search system to improve your search experience.
- We can model your data structure and text analysis
- We can design better queries to satisfy your requirements
- We can tune search components and features to target your objectives
- We can integrate advanced boosting/techniques to improve the quality of results.
Search Results Evaluation in Apache Solr
SEARCH system quality evaluation
Being active researchers in the topic of search results evaluation, we are experts in evaluating the quality of your search results.
- We can evaluate your search results and give you a detail report
- We can integrate your search system to our free tool Rated Ranking Evaluator Enterprise, for an easy collection of user’s feedback and overview of your search system quality
Apache Solr optimization Services
Performance Tuning
Apache Solr is a complex project: many things can be achieved in many ways and it’s often not that easy to choose the best implementation and configuration to minimise the memory footprint, disk and CPU utilisation.
And we know very well that resources = money and exhausting resources can lead to production issues and problems.
At Sease, we know the intricacy of Apache Solr internals and our Information Retrieval specialists can help you minimise your resource consumption and maximise the performance of your system.
Query Response/indexing tim
Speed is vital in any search system and Apache Solr is no different.
Achieving good indexing speed allows for fresher results to be available sooner and fast query response time makes your users happy!
We can guide you toward query optimisation and techniques that can speed up your system.
Apache Solr Feature Integration
Our Apache Solr specialists can help you integrate various features to improve the quality of your search system. Choose from one of these or feel free to ask for others.
spellchecker
faceting
sorting
typeahead
multilingual search
Our Apache Solr exploration and discovery service provides tailored insights and strategic recommendations to optimize your search platform.
We offer expert advice on the architecture, configuration, and usage to enhance user experience and accommodate future needs.
You will get a detailed report, organised to highlight what you have done well and what could have been done better, in a readable and comprehensive format, including configuration/queries examples and corrections.
And the cherry on the cake, we’ll follow up with a call to clarify the report and respond to questions.
Migration to Apache Solr
- Evaluation of pros and cons of migrating from a different search technology to Apache Solr.
- Migration from Elasticsearch, OpenSearch, Vespa and other search technologies to Apache Solr.
Apache Solr Version Upgrade
- Evaluation of pros and cons of upgrading to another version of Solr.
- Upgrade your current Solr version to the latest one
We undertake the technical design and development of complete search systems or specific features, tailoring solutions to meet our client’s project requirements. We aim to deliver efficient and reliable solutions that exceed expectations.
We are domain agnostic: from an enterprise search engine to e-commerce, from Music to TV shows and economics: we have a deep theoretical and practical understanding of information retrieval techniques and have successfully applied them to Solr in many different domains over the years.
Large Language Models can be a powerful allied to power your Solr search.
Our information retrieval experts can guide you through selecting the optimal Large Language Model (LLM) for your project and domain, customizing it to suit your specific needs, and quantifying the enhancements achieved.
- Identify the best Large Language Model for your project
- Customize the LLM and integrate it into your search system
- Evaluate the enhancements achieved with the integration of the Large Language Model.
We have been exploring various LLM integrations such as:
- Natural Language to structured queries (see our talk at Berlin Buzzwords and SMDX)
- Apache Solr Neural Highlighting plugin
There are many ways in which our team of Apache Solr experts can enhance your Solr experience using Large Language Models.
Contact us to find out how to proceed with your project and domain.
Retrieval Augmented Generation opens the door to a completely new user experience where the top-k results of your Solr search are used to power a generative step that provides answers with citations in natural language.
Our team specialize in implementing and evaluating RAG, relying on Apache Solr as your main search backend.
- We evaluate the pros and cons of the implementation of a Retrieval Augmented Generation system into your project
- We integrate and optimize the RAG system into your search engine
We contributed vector-based search to Apache Solr!
Since then we’ve been extending and maintaining the official feature introducing improvements based on the experience we are accumulating with our clients.
From chunking to the choice of the best embedding model for your use case, our Apache Solr expert can help you implement vector-based search with Apache Solr.
Hybrid search techniques allow you to mix traditional keyword search with vector-based search and get the benefits of both worlds.
We are contributing new functionalities in Solr about this (see our talks at Berlin Buzzwords and Community Over Code EU), and our team of Apache Solr experts can help you implement it in your use case.
Learning To Rank adds the ability of integrating Machine Learning models to learn better ways of ranking your search results based on many relevance, users’ and business factors.
We’ve been working hands-on in the Learning To Rank Solr module since 2015 and we were among the founder group (with Bloomberg).
Since then we’ve been actively maintaining and extending the functionality and presenting at conferences many times in this regard (see our blog posts about learning to rank).
We applied Learning To rank in many clients’ projects and we would be delighted to help you in implementing it in your Apache Solr instance.
We’ve been integrating NLP techniques in Solr since 2013, with a focus on Named Entity Recognition and Linking, Query intent classification, etc.
With the advent of modern models such as GliNER, this is still incredibly relevant.
We can help you designing and implementing such integration with your Solr instance.
Many techniques can implemented in Apache Solr to enrich and expand your query and your document for a better recall and user experience.
Our Apache Solr experts can help you integrate learned sparse vector models into your Solr search and get the benefits and explainability of such interesting approaches!
Multi-modal search allows you to search across different media such as text, video, audio and images.
We can help you introduce this functionality in your Solr search engine or improve your current solution.
Search Relevancy Tuning in Apache Solr
Returning relevant search results is hard, but we are up for the challenge!
We’ve been working on improving the relevance of search systems for many years and in many different domains, we explored many techniques, presented approaches at conferences and hosted many pieces of training on the topic.
We can tune the search relevancy of your Apache Solr search system to improve your search experience.
- We can model your data structure and text analysis
- We can design better queries to satisfy your requirements
- We can tune search components and features to target your objectives
- We can integrate advanced boosting/techniques to improve the quality of results.
Search Results Evaluation in Apache Solr
Being active researchers in the topic of search results evaluation, we are experts in evaluating the quality of your search results.
- We can evaluate your search results and give you a detail report
- We can integrate your search system to our free tool Rated Ranking Evaluator Enterprise, for an easy collection of user’s feedback and overview of your search system quality
Apache Solr is a complex project: many things can be achieved in many ways and it’s often not that easy to choose the best implementation and configuration to minimise the memory footprint, disk and CPU utilisation.
And we know very well that resources = money and exhausting resources can lead to production issues and problems.
At Sease, we know the intricacy of Apache Solr internals and our Information Retrieval specialists can help you minimise your resource consumption and maximise the performance of your system.
Speed is vital in any search system and Apache Solr is no different.
Achieving good indexing speed allows for fresher results to be available sooner and fast query response time makes your users happy!
We can guide you toward query optimisation and techniques that can speed up your system.
Our Apache Solr specialists can help you integrate various features to improve the quality of your search system. Choose from one of these or feel free to ask for others.
Spellchecker
Faceting
sorting
typehead
multilingual search
Client's Feedback
Canva solicited Sease’s guidance when we were facing some challenging scale and performance problems with our Solr clusters. Specifically, we had issues with our Solr7 cluster configuration & high latencies on tail queries. They worked with us to understand the key challenges with Canva’s implementation of Solr & its various collections. They also worked with us to understand how we were operating our Solr clusters at the scale of 1000 RPS to deeply investigate any fundamental operational mistakes. As a result of this partnership, they provided a very comprehensive report of suggestions for improvement. These suggestions allowed us to smoothly upgrade the major version of Solr to 8. We have also seen improved stability of the service, with a reduction in latency spikes under heavy load.
Ashwin Ramesh
Director of Engineering @Canva