Senior AI Engineer, Production Agents (Hybrid)

Bausch+Lomb

Bridgewater Township (NJ)

Hybrid

USD 150,000 - 200,000

Full time

14 days+

Get a reply from this employer

Your resume and cover letter, tailored to exactly what they’re hiring for.

Get past ATS filters

Benefits offered by this job

Hybrid work schedule (up to 2 days WFH

Medical, dental, vision benefits

401(k) with company match

Tuition reimbursement

Paid vacation & sick time

Job summary

Bausch + Lomb is seeking a Lead AI Engineer, Agentic AI Solutions, to design and deploy AI-powered tools that support commercial decision-making across brand strategy, customer engagement, market access, and analytics.

The role combines AI engineering with business collaboration to deliver enterprise-scale solutions using modern AI tech, cloud platforms, and large-scale data ecosystems. Hybrid work in Bridgewater, NJ is available with strong governance and compliance expectations.

Qualifications

Bachelor’s or Master’s degree in Computer Science, Artificial Intelligence, Machine Learning, Engineering, or a related field.
5+ years of experience building and deploying production software, AI, or ML solutions, including LLM-powered applications.
Strong Python development skills and software engineering best practices, including testing, CI/CD, APIs, code reviews, and containerization.
Hands-on experience with AI agent frameworks such as LangChain, LangGraph, CrewAI, AutoGen, LlamaIndex, or similar.
Experience with tool calling, structured outputs, function schemas, and enterprise system integrations.
Experience designing and implementing RAG solutions, vector databases, embeddings, and retrieval systems.
Strong SQL and data modeling skills with the ability to define and validate business metrics.
Experience developing frameworks and metrics to evaluate AI/LLM performance.
Experience with cloud and MLOps platforms such as AWS, Azure, Databricks, Snowflake, and orchestration tools.
Experience operating in regulated environments with a focus on governance, compliance, and auditability.
Proven ability to collaborate effectively across technical and business functions.
Experience working in Agile environments using Scrum, SAFe, or similar methodologies.
Strong communication and presentation skills, with the ability to translate complex technical concepts for non-technical audiences.

Responsibilities

Design, build, and operate production Applied agents across the commercial process - brand strategy, customer insights, personalized engagement, and closed-loop feedback composing foundational agents rather than rebuilding them.
Build performance investigation agents that autonomously decompose brand performance shifts across geography, channel, payer, and eye care professional (ECP) segment, returning sourced, reproducible narratives to brand and insights teams.
Build pre-call, post-call, and next-best-action / next-best-engagement agents that assemble ECP briefs from Glimpse (Next Best Action) platform / CRM attach explainable reasoning to every recommendation and write structured data back to CRM.
Build measurement and feedback agents supporting A/B and multivariate test design, readout, and write-back into the planning loop.
Ground all agents in the enterprise semantic layer and commercial knowledge graph so metric definitions, joins, grain, and row- and role-level access are resolved before query generation rather than corrected afterward.
Implement determinism and reproducibility controls, including validate-then-lock specification patterns, so any number that drives a targeting, content, or investment decision returns the same result across runs and is fully auditable.
Integrate agents with enterprise systems; Glimpse (NBA platform) /Veeva or Salesforce, Adobe and internal commercial platforms using Model Context Protocol (MCP), agent-to-agent (A2A) patterns, tool discovery, scoped authentication, and structured outputs.
Build and maintain evaluation harnesses per agent, including golden datasets, calibrated LLM-as-judge scoring, and regression testing, and track agent KPIs covering task completion, grounding, reproducibility, human-override rate, latency, and cost per task.