Forward Deployed Engineer – AI/ML Systems

Role: Forward Deployed Engineer (FDE) – AI-First SystemsLevel: PhD / Post-Doctoral / Equivalent Principal-Level Applied ResearchLocations: San Jose, CA & New JerseyWork Model: Embedded / Hybrid with significant onsite presence and travelInitial Engagement: Approximately 2 weeks/month onsite during the first 6 months, alternating with remote build periodsDomain: Food Manufacturing | Supply Chain | Logistics | AI/ML | OptimizationRole OverviewWe are seeking a highly technical Forward Deployed Engineer (FDE) to design, build, deploy, and operate production AI/ML systems directly within manufacturing and logistics environments.This is not a traditional research, data science, analytics, or advisory consulting role. The ideal candidate must be a hands-on builder who can own solutions end-to-end—from messy operational data and model selection through production deployment, monitoring, troubleshooting, and measurable business outcomes.The engineer will work directly with factory, warehouse, logistics, and business teams, using real-world data from ERP, WMS, IoT sensors, cold-chain telemetry, order systems, procurement platforms, and transaction feeds.Key ResponsibilitiesDesign, build, deploy, and operate production AI/ML systems.Own 2–3 production AI use cases from concept through deployment and optimization.Build end-to-end ML pipelines covering data ingestion, feature engineering, training, evaluation, deployment, and monitoring.Develop real-time and streaming pipelines using Kafka, Flink, Spark, or equivalent technologies.Build low-latency inference services, including sub-100ms P99 use cases.Develop reusable feature-store architectures using Feast or equivalent platforms.Build experimentation frameworks connecting model performance to measurable business outcomes.Develop production LLM/GenAI systems, including RAG, prompt engineering, fine-tuning, and agentic workflows.Apply probabilistic ML, causal inference, reinforcement learning, optimization, and advanced ML techniques to operational problems.Deploy computer vision, anomaly detection, sensor fusion, and Edge AI solutions.Optimize models for edge environments using quantization, pruning, ONNX, and related techniques.Design resilient systems capable of handling intermittent connectivity, sensor failures, and degraded environments.Build scalable lakehouse architectures using Delta Lake, Apache Iceberg, or equivalent technologies.Establish data contracts, data-quality controls, monitoring, and source-level SLAs.Implement secure and auditable AI decision systems.Troubleshoot production issues directly with operators and business teams.Continuously improve deployed systems and translate technical improvements into measurable operational or financial results.Core Technical DomainsCandidates should demonstrate deep expertise in at least 4 of the following areas:Probabilistic ML & Bayesian InferenceBayesian modeling, posterior estimation, variational inference, MCMC, uncertainty quantificationDeep LearningTransformers, attention mechanisms, multimodal models, GNNsReinforcement Learning & BanditsContextual/multi-armed bandits, RLHF, sequential decision systemsCausal Inference & ExperimentationCausal graphs, A/B testing, experimental design, Difference-in-Differences, instrumental variablesMLOps / Production MLModel serving, feature stores, model registries, drift detection, ML CI/CD, monitoringData Engineering & Distributed SystemsKafka, Flink, Spark, lakehouse architecture, distributed pipelines, real-time processingNLP / Generative AILLMs, RAG, fine-tuning, prompt engineering, agentic AIComputer Vision / Sensor FusionProduction CV, anomaly detection, multimodal/sensor integration, edge deploymentOptimization / Operations ResearchMixed-integer programming, stochastic/combinatorial optimization, routing and schedulingHuman-in-the-Loop AIRLHF, annotation pipelines, decision traces, gold-label constructionRequired QualificationsPhD in Computer Science, Statistics, Applied Mathematics, Operations Research, Electrical Engineering, or a related quantitative discipline.OR equivalent research depth, such as:Post-doctoral research experienceFirst-author publications at top-tier venues5+ years of principal-level applied research at a research lab or AI-first companyAdditional requirements:Proven production ML experience with models deployed, used by real users, and monitored in production.Strong Python skills with technologies such as NumPy, PyTorch/JAX, scikit-learn, Pandas, and SQLAlchemy.Experience with at least one compiled/JVM language: Java, Scala, Go, or Rust.Experience with AWS SageMaker, GCP Vertex AI, or Azure ML.Experience deploying and serving models on Kubernetes.Strong understanding of distributed systems and real-time production architectures.Ability to work with incomplete, noisy, and heterogeneous operational data.Ability to select the appropriate technical approach rather than applying ML unnecessarily.Strong communication skills with technical, operational, and business stakeholders.Comfortable working in rapidly changing environments with evolving requirements.Strongly PreferredPrevious Forward Deployed Engineer, Embedded ML Engineer, or AI-in-Residence experience.Applied AI experience in manufacturing, logistics, supply chain, food/agriculture, healthcare, energy, or government.Research or publications in probabilistic forecasting, causal ML, optimization, reinforcement learning, or human-in-the-loop systems.Experience with IoT, sensor data, and Edge ML.Experience with logistics, routing, or scheduling optimization.Experience building regulated or auditable AI systems.Proven ability to connect technical/model improvements to measurable business or financial outcomes.Work-Style RequirementsThis is a hands-on engineering role, not an advisory position.The successful candidate must be comfortable:Writing production codeBuilding and debugging data pipelinesDeploying and monitoring ML modelsWorking directly in manufacturing and warehouse environmentsTroubleshooting model drift, sensor failures, and data-quality issuesRapidly shipping production fixesTaking end-to-end ownership of assigned AI use casesCandidates focused primarily on research, prototyping, product analytics, roadmap management, or advisory consulting are not a fit.Travel & Embedded OperationsDuring the initial 6-month deployment period, the FDE will work in approximately 2-week embedded sprint cycles, with onsite travel to San Jose and New Jersey facilities approximately 2 weeks per month.Candidates must be comfortable working directly with operational teams at client facilities.#Hiring #AI #MachineLearning #ArtificialIntelligence #AppliedAI #MLEngineer #MLOps #GenAI #LLM #DeepLearning #Python #Kubernetes #ComputerVision #Optimization #SupplyChain #Logistics #Manufacturing #EdgeAI #SanJoseJobs #NewJerseyJobs