Sr Software Engineer - Coding Agents

Lexsi Labs; Mumbai; Hybrid; INR 1,800,000 - 2,400,000; Full time

Job summary

Lexsi Labs in Mumbai invites a developer tooling engineer to strengthen our autonomous-system tooling stack. You will work on the agent architecture, retrieval, editing, builds, tests, static analysis and version control within a customer environment.

This role owns the coding agent, ensuring changes are verified before humans see them, with large legacy codebases and limited internet egress. You will collaborate with research, focus on performance, safety and interpretability, and build

Qualifications

• Deep experience with developer tooling and IDEs, editor and IDE internals.
• Experience with language servers, LSP and code intelligence tooling.
• Strong knowledge of build systems, CI/CD and hermetic builds.
• Familiarity with observability tooling and debugging complex systems.
• Experience building evaluation datasets or frameworks for AI systems is a plus.

Responsibilities

• Own the agent tooling surface including retrieval, editing, builds, tests, static analysis, and version control.
• Verify changes automatically so a human only reviews after correctness is established.
• Support long-horizon execution across multi-refactor migrations with checkpointing and recovery.
• Define authorization and rollback models for agent actions.
• Integrate with engineering stacks and observability tooling to debug and audit behavior.
• Develop evaluation suites from real repository tasks and tune behavior accordingly.

Skills

Developer tooling

IDEs

Language servers

LSP

Code intelligence

Refactoring engines

CI/CD practices

Observability

Performance & safety

Evaluation datasets

Tools

Bazel

Gradle

Maven

Make

Jenkins

GitLab CI

GitHub Actions

Job description

LexsiLabsistheleadingfrontierAIlabfocusedonbuilding aligned,interpretable,andsafesuperintelligentsystems .Whilethatisthevision,themissiontobuildsafetyawareautonomoussystemintheextremenearterm.OurresearchworkspansaroundareaslikeAIalignmentmethodologies,interpretability ledsystemdesign,andfoundationalmodelresearchacrossstructured,tabular,andnewautonomoussystemdesigns.Wepublishedabout25+papersinthepast15monthsacrossleadingconferencesICLR,ICML,WWW,IJCNN,MICCAI,Euripsetc.OurlabsarelocatedinIndia(Mumbai&remote),Paris,&London.

Weoperatewithaflatstructure,highautonomy,andastrongbiastowardengineerswhotakefullownershipofwhattheybuild,fromarchitecturetoproductionbehavior.

TheRole

Ourcurrentsprintonbuildingautonomoussystemforcomplexproblems,acrosssoftwareengineering,datascience,andAIresearchinvolvesbuildingharness,executionsubstrateandevaluationsystem,andeachisdesignedtoruninsideacustomer'senvironmentratherthanours.

Thisroleownsthecodingagent:asystemthattakesanobjective,worksdirectlyonarepository,verifiesitsownchanges,andproducesarecordofwhatitdid.Thecodebasesarelarge,old,thinlytestedandload-bearing,andtheagentrunsinsidethecustomer'snetwork,usuallywithoutinternetegress.

What'llyouworkon:

• Theagentarchitectureanditstoolsurfaceforsoftwarework,includingretrieval,editing,builds,tests,staticanalysisandversioncontrol.
• Verification,sothatachangeisestablishedascorrectbeforeahumanseesit.
• Long-horizonexecutionacrossmulti-filerefactorsandmigrations,withcheckpointing,recoveryandself-correction.
• Theauthorisationandrollbackmodelgoverningwhattheagentmaydoandhowitisundone.
• Integrationswiththeengineeringestatetheagentoperatesin,andtheobservabilitytoolingusedtodebugandauditit.
• Evaluationsuitesbuiltfromrealrepositorytasks,andthebehaviortuningthatfollowsfromthem.

Youwillworkcloselywithourresearchteamonevaluationdesign,post-trainingandinterpretabilityofagentbehavior.

WhatWeAreLookingFor

ThisisadevelopertoolingroleasmuchasanAIrole.Theagentisonlyascapableasthetoolsandintegrationsitisgiven,sodepthheremattersmoretousthananythingelse.

• DeepexperiencewithdevelopertoolingandIDEs.EditorandIDEinternalsorextensions,languageserversandLSP,codeintelligence,refactoringengines.Youhavebuiltthingsengineersuseeveryday.
• Codeintelligenceandprogramanalysis.ASTsandtree-sitter,staticanddataflowanalysis,symbolindexingandcodesearch,callgraphanddependencyresolution,codemodsandautomatedmigrationtooling.
• BuildsystemsandCI/CD.Bazel,Gradle,Maven,Makeandsimilar,incrementalandhermeticbuilds,artifactanddependencymanagement,pipelinedesignacrossJenkins,GitLab,GitHubActionsorequivalents,includingself-hostedandofflinedeployments.
• Observability.Distributedtracing,structuredlogging,OpenTelemetry,andbuildingthereplayandinspectiontoolingthatmakesnon-deterministicsystemsdebuggable.
• Integrationsengineering.GithostingAPIs,ticketingandworkflowsystems,enterpriseauth,andtherealityofworkingagainstself-hostedsoftwareseveralversionsbehindcurrent.
• Agenticsystemsinproduction,notprototypes.Youknowthecurrentframeworklandscape(ReAct-styleagents,LangGraph,LangChain,SemanticKernel)andwhereeachonestopsworking.
• Youreasonaboutperformance,reliability,cost,safetyandinterpretabilityasoneconnectedsetofconstraints.
• Youmakegoodcallsinthegrayarea,andyouarecomfortablewhenproblemsarelooselyspecifiedandownershipisassumedratherthanassigned.
• ExperiencebuildingevaluationdatasetsorframeworksforAIsystemsisaplus.
• Compilerbackground,largemonorepowork,orlegacymodernisationprogramsareaplus.
• Priorexposuretoalignment,interpretabilityorsafetytoolinginproductionsettingsisaplus.

Wemovequicklyandexpectcandidatestodothesame.Wevaluesubstanceoverpolishandexecutionoverrhetoric.