Staff AI Engineer

Meeru AI Inc; Pleasanton (CA); On-site; USD 180,000 - 260,000; Full time

Job summary

Meeru AI Inc is building an AI layer with grounded, verifiable outputs. ThisStaff AI Engineer role is senior, setting technical direction for ML/LLM and agent systems, closely partnering with evaluation to ensure quality.

The role is onsite in Pleasanton, CA, supporting deployments where data remains in the customer's environment. You will own the AI layer, ML models, LLM/agent systems, and per-customer models with strict anti-leakage.

Qualifications

• Staff-level applied ML/AI with production delivery experience.
• Hands-on depth in time-series modeling and anomaly/outlier detection.
• Experience with gradient-boosted models (LightGBM/XGBoost) in Python.

Responsibilities

• Own the AI and intelligence layer and the discipline that keeps it honest.
• Own the machine-learning models that learn each customer’s patterns.
• Own the LLM and agentic systems that produce grounded outputs.
• Build per-customer models with anti-leakage discipline.
• Set a high bar for evaluation, grounding, and hallucination control.
• Provide technical direction for the AI layer and collaborate with evaluation.
• Turn user feedback into durable improvements.

Skills

Staff-level ML/AI

Time-series modeling

Anomaly detection

Gradient-boosted models

LightGBM

XGBoost

Python

scikit-learn

pandas

Polars

LangGraph

RAG

SQL

Snowflake

PostgreSQL

LangChain

Tools

LightGBM

XGBoost

Python

scikit-learn

pandas

Polars

SQL

Snowflake

PostgreSQL

LangGraph

LangChain

Job description

Meeru AI Inc is building an AI and intelligence layer where outputs must be grounded, traceable to verified data, and evaluated for hallucination control. This Staff AI Engineer role is a hands-on, senior position that sets the technical direction for ML, LLM, and agent systems while partnering with evaluation to make quality measurable. The work is delivered onsite in Pleasanton, CA, supporting deployments that keep customer data in their environment.

What you will own

• Own the AI and intelligence layer and the discipline that keeps it honest.
• Own the machine-learning models that learn each customer’s patterns.
• Own the LLM and agentic systems that produce grounded natural-language output.
• Build per-customer models with strict anti-leakage discipline so models do not expose the signals they detect.
• Set a high bar for evaluation, grounding, and hallucination control across the layer.
• Provide technical direction for the AI layer.
• Partner closely with the evaluation engineer to make quality measurable.
• Turn user feedback into durable improvements over time.

What you bring

• Staff-level applied ML/AI experience delivering models in production.
• Strong classical ML with hands-on depth in time-series modeling, anomaly and outlier detection, and gradient-boosted models such as LightGBM or XGBoost, using Python, scikit-learn, and pandas or Polars.
• Modern LLM and agent patterns, including retrieval-augmented generation (RAG), tool-using agents, structured (schema-validated) output, and orchestration with LangGraph or a similar framework.
• Rigor around evaluation, grounding, and hallucination control, including measuring whether outputs are faithful to source data and building guardrails that prevent unsupported statements from reaching users.
• Clear discipline on data leakage and honest evaluation, designing features, splits, and validation so training and scoring reflect what is available in real use.
• Clear thinking about what ML can and cannot claim, including communicating to non-ML colleagues where outputs are evidence vs. signals.
• Real software engineering: tested, production-grade Python services; comfort with SQL against Snowflake or PostgreSQL; and shipping containerized software.
• Technical leadership through mentoring and close collaboration with backend, data, product, and evaluation teams to convert feedback into lasting improvements.

Why this role is different

• Grounding is non-negotiable: no fabricated or unsupported output, ever.
• Faithfulness is enforced with evaluation as a gate, not just a tuning target.
• Separation of concerns prioritizes ML surfaces and ensures generated text states what is supported by verified data or confirmed by a human.
• Per-customer intelligence models learn business behavior with strict anti-leakage discipline.
• Customer-cloud deployments including managed LLMs in-VPC so customer data does not leave their environment.

Nice to have

• Finance or accounting domain experience.
• Feature-attribution literacy, including experience with SHAP.
• Human-in-the-loop systems, including review, confirmation, or correction workflows fed back into models.
• Experience running managed LLMs inside customer clouds in a VPC (AWS Bedrock, Google Vertex AI, or Azure OpenAI).
• Sparse and zero-inflated data experience.Multi-deployment enterprise software exposure, including multi-tenant SaaS, single-tenant SaaS, and on-premises across AWS, Azure, and GCP.

Technologies

Python, scikit-learn, pandas, Polars, LightGBM, XGBoost, LangGraph, RAG, SQL, Snowflake, PostgreSQL, LangChain