{"id":1220912,"url":"https://alion.io/job/hrs-agentic-ai-automation-engineer","title":"Agentic AI Automation Engineer","company":{"id":678572,"name":"HRS","domain":"hrs.com","url":"https://alion.io/company/hrs-2","size_band":null,"is_staffing_agency":false,"employer_type":"direct","is_intermediary":false,"listed_via":null,"ats_vendor":"Ashby","truth_index":null},"role":"AI/ML","role_family":"AI/ML","seniority":"middle","employment_type":"full_time","work_mode":"on_site","remote_scope":null,"remote_scope_basis":null,"remote_working_hours":null,"hiring_geo_confidence":"structured","locations":["Mohali, India"],"countries":["IN"],"hiring_countries":[],"hiring_countries_total":0,"salary":null,"salary_estimate":{"min_usd":23000,"max_usd":58000,"period":"year","method":"role_seniority_country_remote_cell","sample_n":22},"experience_years_min":3,"visa_sponsorship":false,"relocation_package":false,"has_equity":false,"technologies":[{"name":"Agile","optional":false},{"name":"AI Agents","optional":false},{"name":"AIOps","optional":false},{"name":"Amazon EKS","optional":false},{"name":"AWS","optional":false},{"name":"AWS Bedrock","optional":false},{"name":"AWS Bedrock AgentCore","optional":false},{"name":"CI/CD","optional":false},{"name":"Claude","optional":false},{"name":"CrewAI","optional":false},{"name":"DSPy","optional":false},{"name":"GitHub","optional":false},{"name":"Google ADK","optional":false},{"name":"Hallucination","optional":false},{"name":"Human-in-the-Loop","optional":false},{"name":"LangChain","optional":false},{"name":"LangGraph","optional":false},{"name":"LLM","optional":false},{"name":"LLM Guardrails","optional":false},{"name":"n8n","optional":false},{"name":"New Relic","optional":false},{"name":"Python","optional":false},{"name":"RAG","optional":false},{"name":"Terraform","optional":false},{"name":"Kubernetes","optional":true},{"name":"Milvus","optional":true},{"name":"pgvector","optional":true},{"name":"PostgreSQL","optional":true},{"name":"UiPath","optional":true}],"status":"live","first_seen_at":"2026-09-25T10:52:44Z","employer_posted_date":"2026-09-25","last_verified_at":"2026-09-30T16:26:41Z","board_verified":true,"closed_at":null,"days_open":5,"trust":{"level":"ok","repost_count":null,"flags":[],"days_open":5},"description":"Mid-Level AI Engineer - Agentic AI\nAI-Workflow Programme | Mohali, On-Site\nPOSITION\nWe are seeking a Mid-Level AI Agentic Engineer to join the AI-Workflow programme and\nbuild the autonomous crew systems that augment HRS operations across Finance,\nControlling, Operations, Customer Service, Customer Experience, and HR. This is not a\nresearch role or a prototype environment - you will be building production AI crews that\nhandle live operational workflows for real departments, with real outcomes measured\nfrom day one.\nYou will work within a \"crews building crews\" model: a platform of seven build agents (PM,\nArchitect, Automation, QA, SRE, Documentation, Observability) scaffolds, tests, and\ndocuments the operational crews you build. Your job is to close the gap between agent\nscaffolded output and production-ready code - working directly with the Tech Lead, the\nPM, and the build agent platform to deliver tested, instrumented, and documented crews\nwithin a 10-day delivery lifecycle.\nThe mission is workforce augmentation. AI handles the volume. Humans handle the\njudgement. Every crew you build encodes that principle in every escalation boundary,\nevery guardrail, and every human-in-the-loop gate.\nCHALLENGE\nCrew Development & Implementation\nBuild operational AI crews from structured To-Be process descriptions using DSPy\ntyped signatures with assertion guards and agent workflow orchestration patterns\nsuch as state machines, human-in-the-loop checkpoints, and resumable execution\nImplement N8N workflow automation and JSON integration connectors linking\ncrews to operational systems including Zammad, Genesys, and enterprise back\noffice platforms\nWork directly with the Automation Agent to scaffold DSPy modules and agent\nworkflows - extending and improving generated output, not accepting it verbatim\nDesign and encode specific, testable escalation boundaries for every crew before\nshadow deployment - grounded in real process context, not generic confidence\nthresholds\nDeliver every crew with 100% unit test coverage, a complete runbook, and New\nRelic instrumentation live before go-live - these are deployment gates, not\naspirational standards\nContribute reusable patterns to the shared crew library and peer-review modules\nbuilt by other engineers on the team\nTechnical Execution & Quality\nImplement agent memory management using explicit typed state, structured\ncontext handling, and clear handoff boundaries to prevent context degradation\nacross multi-step operational workflows\nApply three-layer output validation - DSPy assertions, output validators, and policy\nenforcer - on every crew module before merge\nBuild and validate test suites covering non-deterministic edge cases and failure\nmodes - not just happy paths - using the QA Agent's generated baseline as a\nstarting point\nIntegrate crews with AWS Bedrock model routing (Claude Haiku/Sonnet) and work\nwithin the EKS and Terraform IaC stack managed by DevOps\nMaintain guardrails configuration for every crew - escalation triggers, human\napproval gates, and policy enforcement - encoded in config before any crew\nenters shadow deployment\nParticipate in weekly DSPy evaluation cycles against gold-standard baselines to\nvalidate crew output quality and flag drift\nObservability & Production Operations\nInstrument every crew with New Relic metrics from day one: throughput, error rate,\nlatency, escalation rate, and cost per task - observability is a deployment\nprerequisite, not an afterthought\nActively diagnose and resolve production failure modes: memory drift across multi\nstep workflows, hallucination under low-confidence RAG retrieval, context\ndegradation in long-running state machines, and prompt injection via untrusted\nintegration inputs\nUse post-deployment observability data to identify improvement candidates and\nraise them in RAID - closing the feedback loop the Observability Agent depends on\nContribute to the continuous improvement cycle: every crew in production is a\nmeasurement and improvement loop, not a delivery milestone\nCollaboration & Build Platform\nWork within the 10-day delivery lifecycle - Request → Discovery → Design →\nDevelopment → QA → CI/CD → Monitoring → Continuous Improvement - delivering\nto standard at each stage\nCollaborate with the PM during Discovery to assess process automation feasibility\nusing FUDV scoring - frequency, uniformity, digitisation, volume - and push back\ncredibly where AI reliability or data quality is not there yet\nContribute to Thursday technical reviews and Friday retrospectives with substantive\ninput - not status updates but engineering judgment\nUse the build agent platform as a personal productivity multiplier - flag platform\ngaps via RAID rather than working around them silently\nFOR THIS EXCITING MISSION YOU ARE EQUIPPED WITH…\nAgentic AI Technical Skills\n3-5 years of experience in AI/ML development with 1+ years in agentic AI or\nadvanced LLM applications shipped to a production environment - not prototype\nor hackathon experience\nHands-on experience with DSPy typed signatures and assertion guards - not just\nLangChain familiarity\nPractical exposure to at least one agent framework or platform such as LangGraph,\nGoogle ADK, Amazon Bedrock AgentCore, LangChain, CrewAI, or equivalent; the\nrole values transferable agentic engineering patterns over any single required\nframework\nExperience building and committing N8N workflow automation in a production\ncodebase\nDemonstrated ability to design specific, testable escalation boundaries in a live\noperational AI system\nCan show their work - a GitHub profile, a shipped system, or a concrete\nbefore/after on a workflow they automated carries more weight than academic\ncredentials\nAI Engineering Capabilities\nStrong Python programming skills with AI/ML libraries and practical agentic\nengineering patterns; able to work across frameworks when needed, with exposure\nto at least one of LangGraph, Google ADK, Amazon Bedrock AgentCore, LangChain,\nCrewAI, or equivalent\nProduction experience with AWS Bedrock or equivalent cloud-based LLM routing\nand model management\nKnowledge of vector databases, embedding systems, and retrieval-augmented\ngeneration - including retrieval quality assessment and hallucination mitigation\nUnderstanding of MLOps and AIOps practices: CI/CD for AI systems, evaluation\nharnesses, and gold-standard baseline testing\nHands-on experience with New Relic or equivalent observability tooling for\nproduction AI systems - metric design, dashboard instrumentation, and anomaly\ndiagnosis\nFamiliarity with containerisation, EKS, and Terraform IaC sufficient to work within a\nDevOps-managed infrastructure without creating integration delays\nDevelopment & Process Skills\nTest-driven development for non-deterministic systems - 100% unit test coverage\nbefore merge is a non-negotiable standard in this team\nExperience with agile delivery in a timeboxed sprint model - able to take a\nstructured process description from design to shadow deployment within a 10-day\nlifecycle\nStrong code documentation discipline - every module peer-handoff ready, every\nrunbook complete during build, every decision traceable in RAID\nAbility to work within an architecture set by a Tech Lead - executing with full\nownership and quality pride within defined guardrails, escalating cleanly when\nconstraints need revisiting\nProfessional Skills\nWrites to be understood, not to be impressive - RAID entries a director can triage,\nrunbooks a department SME can follow, code a peer can extend without asking the\nauthor\nCalm under non-determinism - diagnoses production failures methodically using\nobservability data rather than thrashing or going silent\nDog-food mentality - uses the build agent platform to accelerate their own work\nand actively contributes to improving it\nMission-driven - understands that the goal is workforce augmentation, not\nautomation for its own sake, and builds every crew with that principle at its centre\nPreferred Experience\nDomain experience in at least one of: Finance, Controlling, HR, Operations,\nCustomer Service, or Customer Experience - brings escalation boundary instinct\nthat no intake card can fully replicate\nExperience with enterprise system integrations: Zammad, Genesys, UiPath, or\nequivalent CRM/CX/RPA platforms\nFamiliarity with ChromaDB/Milvus/pgVector or equivalent vector store for RAG\npipeline development\nExperience contributing to a shared pattern library or internal engineering\nknowledge base in a multi-engineer AI team","description_format":"text","description_chars":8458,"description_truncated":false,"requirements":{"experience_years_min":3,"management_years_min":null,"team_size_min":null,"manages_managers":false,"education":null,"security_clearance":false,"languages":[]},"benefits":[],"hiring_locations":[],"hiring_excludes":[],"relocation_offered":false,"industries":[],"lifecycle":[{"event":"open","at":"2026-09-25T11:56:29Z"}],"liveness":{"score":85,"band":"hot","label":"Hiring now","p_open":1,"p_active":0.853,"p_room":1,"age_days":4,"expected_fill_days":25,"reasons":["conf:8","velocity","win:early"],"computed_at":"2026-09-30T05:45:00Z"},"pay":null,"html_url":"https://alion.io/job/hrs-agentic-ai-automation-engineer","json_url":"https://alion.io/job/hrs-agentic-ai-automation-engineer.json","meta":{"generated_at":"2026-10-01T02:05:34Z","cache_seconds":300,"methodology":"https://alion.io/methodology","terms":"https://alion.io/terms","contact":"https://alion.io/contact","api":"https://alion.io/developers","usage":{"tier":"crawler","counted_by":"address","units_charged":1,"used_today":1745,"day_limit":5000,"remaining_today":3255,"minute_limit":60,"resets_at":"2026-10-02T00:00:00Z"}}}