Lead AI Data Scientist

Xcede; Greater London; Hybrid; GBP 120,000 - 170,000; Full time

Job summary

Xcede partners with a PE-backed legal services group to roll out a centralized AI platform across the UK. You’ll be the first data scientist, building the function and setting its direction.

You’ll prototype LLM solutions, take models to production, map data across matter, client and billing systems, and define success metrics to demonstrate productivity gains. This London-based hybrid role offers real influence as the AI program scales across the group.

Qualifications

• 5+ years as a data scientist with leadership experience.
• Track record of moving models from prototype into production.
• Hands-on experience building and evaluating LLM solutions, including prompt design, RAG and structured evaluation.
• Strong Python and SQL, plus experience on a cloud data platform (Databricks, Azure or Snowflake).
• Experience defining success metrics and showing measurable business impact.

Responsibilities

• Collaborate with lawyers, subject-matter experts, business analysts and engineers to identify where AI can impact legal workflows.
• Prototype LLM-based solutions and test them using human-in-the-loop review and LLM-as-judge evaluation, then guide successful ones into production.
• Map data across matter, client, document, time-recording and billing systems and help define a target data model.
• Set baselines and success metrics, and build before-and-after analyses and dashboards to show impact on productivity, quality, cost and margin.
• Monitor models and AI-enabled workflows in production for quality degradation and drift.

Skills

Data science leadership

Production ML/AI deployment

Python

SQL

Stakeholder management

Tools

Databricks

Azure

Snowflake

Job description

LeadAIDataScientist

BuildthedatasciencefunctionataPE-backedlegalservicesgroupgoingall-inonAI

Ourclientisafast-growing,privateequity-backedlegalservicesgroup.They'vegrownthroughasuccessfulbuy-and-buildstrategyandnowbringtogetherseveralleadingregionallawfirmsacrosstheUK.

They'vejustsecuredboardapprovaltorolloutacentralisedAIplatformacrossthewholegroup.AIadoptionisalreadyhigh(around95%oflawyersuseit),butuseisn'tjoinedupacrossthefirms.Thenextphaseismakingitstructured,measurableandscalable.Thisisanewlycreatedroletoleaddatascienceacrossthatprogramme.

Whythisrole?

• Foundingrole:you'llbethefirstdatascientistandwillhire,withroomtogrowthefunctionfromthere.
• Modelsthatship:thisisaPE-backedbusinesswithamandatetodeliver,soyourworkgoesintoproductionratherthansittinginanotebook.
• Realinfluence:you'llbeattheheartofthebusiness,shapinghowanentiregroupusesAI,notoneofhundredsofdatascientists.
• Legalisaheadofthecurve:itsoneofthefewindustrieswhereAIhasmovedfromexperimenttoeverydaypractice.Whatyoudoherewillplayoutinothersectorslater,andyou'llhavedoneitfirst.

Whatyou'lldo

• Workwithlawyers,subjectmatterexperts,businessanalystsandengineerstofindwhereAIandautomationcanhavethebiggestimpactinlegalworkflows.
• PrototypeLLM-basedsolutionsandtestthemusinghuman-in-the-loopreviewandLLM-as-judgeevaluation,thenguidethesuccessfulonesintoproduction.
• Mapdataacrossmatter,client,document,time-recordingandbillingsystems,andhelpdefineatargetdatamodel.
• Setbaselinesandsuccessmetrics,andbuildbefore-and-afteranalysisanddashboardstoshowimpactonproductivity,quality,costandmargin.
• MonitormodelsandAI-enabledworkflowsinproductionforqualitydegradationanddrift.

Whatwe'relookingfor

• 5+yearsasadatascientist,includingtimeleadingorsettingdirectionforDSwork.
• Atrackrecordoftakingmodelsfromprototypeintoproductionandkeepingthemrunning.
• Hands-onexperiencebuildingandevaluatingLLMsolutions,includingpromptdesign,RAGandstructuredevaluation.Exposuretoagents,MCPorcontextengineeringisaplus.
• StrongPythonandSQL,plusexperienceonaclouddataplatform(Databricks,AzureorSnowflake).
• Experiencedefiningsuccessmetricsandshowingmeasurablebusinessimpact.
• Excellentstakeholderskills,withtheabilitytoturnworkflowsintotestableAIproposalsandexplaintrade-offstosenior,non-technicalaudiences.
• Experiencementoringormanagingothers,oraclearambitiontomoveintopeopleleadership.

Nicetohave

• Professionalservices,legalorlegaltechexperience.
• Backgroundinaregulated,data-drivenindustrysuchasfinancialservicesorinsurance.
• PE-backedorinvestor-ownedenvironmentexperience.
• Familiaritywithlegalplatforms(ShareDo,Clio,Aderant)orlegalAItools(Jylo,LexisProtégé,Copilot).
• Awarenessofconfidentiality,privilegeanddata-protectionconsiderationswhenapplyingAItoclientdata.

Thisisafull-timepositionwithahybridset-upfromtheircentralLondonoffice.Unfortunatelyvisasponsorshipandrelocationsupportarenotavailableforthisrole.