AI research engineer

Tricefal; Paris; Sur place; EUR 90 000 - 130 000; Plein temps

Résumé du poste

A European deep-tech startup is seeking an AI Research Engineer to work across the full lifecycle of advanced machine learning systems focused on robustness and real-world impact. Responsibilities include designing neural representations, implementing semantic vectorization techniques, and ensuring alignment between embeddings and code structure. Candidates should have strong software engineering fundamentals and proficiency in Python, along with experience in AI model development. The role is based in Paris and offers a collaborative environment.

Qualifications

• Strong fundamentals of software engineering.
• Strong knowledge of Python and its ML frameworks.
• Experience with the full lifecycle of AI model development.
• Developing new ML methods, algorithms, and models.

Responsabilités

• Design, train, and iterate on neural representations optimized for similarity search.
• Implement semantic vectorization techniques for high-value content.
• Improve content polarization toward precise software constructs.
• Ensure alignment between semantic embeddings, code structure, and runtime behavior.
• Continuously improve precision and recall in large-scale vector search systems.
• Develop evaluation protocols and benchmarks for similarity, relevance and contextual accuracy.

Connaissances

Software engineering fundamentals

Python

AI model development lifecycle

ML methods and algorithms

Description du poste

XxxxxxxisanEuropeandeep-techstartupbuildinganewinfrastructurelayerfor AI-powered softwaresystems.

Our corebeliefissimple,AI doesnotprimarilysufferfromalackof models,itsuffersfroma lackof reliable,structuredandgovernedcontext.

Today’sLLMsoperateon limited,staticcontextwindows,oftendisconnectedfromreal-world softwaresystems.Thisresultsinhallucinations,brittleautomation,technicaldebtamplification andpreventstrueAI autonomy.

Wearebuildingacontextintelligenceplatformfor codeandsoftwaresystems,designedto aggregate,structureandexposestatic,historicalanddynamiccontext,comingfromsourcecode, architectures,businesslogic,runtimebehavior,logs,incidents,supportticketsandoperational signals.

Thiscontextbecomesinfrastructure,consumableby humansandby AI systems. Our technicalplayground

AtXxxxxxx,AI isnotaproductfeature. Itisthecoresystem.

Our work spansacrosshigh-impact,research-gradetopics,including:

• Advanced semantic representations of source code and large-scale software systems
• Alignment between human intent, actual source code and runtime behavior
• Agent orchestration powered by governed, high-fidelity context
• Scaling AI systems in mission-critical enterprise environments
• Xxxxxxx is an agnostic to any AI Models service provider and any vector database provider

Weoperateattheintersectionof AI research,softwareengineering,systemsdesignandreal- worldproductionconstraints.

KeyResponsibilities

Webelievethatthebestresearchisdrivenby strongengineeringfoundations.As partof theAI researchteam,you willoperateacrossthefulllifecycleof advancedmachinelearningsystems, fromideationtoproductiondeployment,withastrongemphasison robustness,precisionand real-worldimpact.Your responsibilitiesinclude:

Researchandmodeling

• Design,trainanditerateon neuralrepresentationsoptimizedfor similaritysearch, includingdense,hybridandmulti-vectorapproaches
• Researchandimplementsemanticvectorizationtechniquesfor high-value,high-signal content,includingstructureddata,unstructureddataandcode-relatedartifacts
• Explorenovelembeddingstrategiesthatcapturesemanticintent,structureandbehavior, notjustsurface-levelsimilarity
• Investigatetrade-offsbetweenembeddingdimensionality,expressiveness,latencyand cost
• Continuouslyimproveprecisionandrecallinlarge-scalevectorsearchsystems
• Developevaluationprotocolsandbenchmarksfor similarity,relevanceandcontextual accuracy

Contextualizationof dataintocode-centricrepresentations

• Improvecontentpolarization,aligninggenericdatatowardprecisesoftwareconstructs suchasmodules,functions,services,domainsor businesscomponents
• Designrepresentationsthatenablecontext-awarereasoningfor AI agentsoperatingon Codebases
• Ensurestrongalignmentbetweensemanticembeddings,codestructureandruntime behavior
• Buildresearchcodethatsupportsfastiteration,reproducibilityandcollaborative experimentation
• Own modelperformanceend-to-end,evaluationquality,trainingefficiency,inference latencyandscalability
• Partnercloselywithproductandplatformteamstodeployresearchoutputsinto productionsystems
• Contributetointernaltools,frameworksandbestpracticesthataccelerateresearch velocity
• Shareinsights,experimentsandfailuresclearlywiththeteam
• Staycurrentwiththestateoftheartinrepresentationlearning,retrievalsystemsand appliedAI research

AtXxxxxxx,researchisnotisolatedfromreality.Youareexpectedtothinkdeeply,experiment boldlyandshipresponsibly,withyourworkdirectlyshapinghow AI systemsunderstand, navigateandreasonovercomplexsoftwaresystemsatscale.

RequiredQualifications

• Strongfundamentalsof softwareengineering
• Strongknowledgeof Python,anditsMLframeworks
• Experiencewith:
  • Fulllifecycleof AI modeldevelopment
  • DevelopingnewMLmethods,algorithmsandmodels