For employers

Hire this AI Trainer

Sign in or create an account to invite AI Trainers to your job.

LoginCreate account
Invite to Job
J
Juliana

Juliana

Agency

Team Leader, Computational Linguistics & AI Evaluation — Google (via Trilyon)

USA flagSan Diego, Usa
www.marchandai.com/

Key Skills

Software

LabelboxLabelbox
Surge AISurge AI
Data Annotation TechData Annotation Tech
MindriftMindrift
Internal/Proprietary Tooling
MercorMercor
TelusTelus
OneFormaOneForma
AppenAppen
DatatureDatature
Other
RoboflowRoboflow
Google Cloud Vertex AIGoogle Cloud Vertex AI
Don't disclose

Top Subject Matter

Enterprise AI agent evaluation and quality assurance (multilingual/vendor compliance)
LLM evaluation
instruction tuning

Top Data Types

TextText
VideoVideo
AudioAudio
DocumentDocument

Top Task Types

Red TeamingRed Teaming
SegmentationSegmentation
Prompt + Response Writing (SFT)Prompt + Response Writing (SFT)
Function CallingFunction Calling
Evaluation/RatingEvaluation/Rating
Entity (NER) ClassificationEntity (NER) Classification
TrackingTracking
ClassificationClassification

Company Overview

Team Leader, Computational Linguistics & AI Evaluation — Google (via Trilyon). Brings 6+ years of professional experience across legal operations, contract review, compliance, and structured analysis. Core strengths include Google Cloud Vertex AI, Internal, and Proprietary Tooling. Education includes Master of Fine Arts, Academy of Art University and Bachelor of Fine Arts, Charles Chaplin Communication Institute. AI-training focus includes data types such as Text and labeling workflows including Evaluation, Rating, and Entity (NER) Classification.

Security

Security Overview

Marchand AI LLC maintains rigorous security and privacy standards appropriate to a boutique AI consultancy handling sensitive client data across regulated industries. Access & Clearance: Principal consultant holds an active Title III Homeland Security/DEA security clearance, establishing a foundational standard of trust and discretion across all engagements. Data Handling: Client data is handled on a strict need-to-know basis. All subcontractors and independent consultants engaged by Marchand AI are bound by confidentiality clauses and non-disclosure agreements prior to project commencement. Regulatory Awareness: Operations are conducted in alignment with HIPAA data privacy principles and FERPA requirements for engagements involving student records or educational institutions. Confidentiality by Design: All client deliverables, workflows, and proprietary methodologies are treated as strictly confidential. Sample frameworks shared externally are fully anonymized. Non-Competing Engagements: Marchand AI exclusively serves non-competing industries to eliminate conflicts of interest, ensuring full client confidentiality across the portfolio. Contracts: All client engagements are governed by a signed Professional Services Agreement including explicit confidentiality, IP ownership, and data privacy provisions.

Security Credentials

HIPPA

Labeling Experience

Google Cloud Vertex AI

Team Lead, Computational Linguistics and AI Evaluation - Google

Google Cloud Vertex AIGoogle Cloud Vertex AITextTextTrackingTrackingRed TeamingRed Teaming

Led a focused team building enterprise-scale Gemini AI automation infrastructure from scratch for a multi-company vendor network. Designed and deployed automated agent workflows for guidelines management, evaluation rubric generation, vendor training delivery, and quality control while eliminating manual processing. Built and tested RAG pipelines integrated with Gemini models to improve multilingual knowledge retrieval accuracy and consistency across workflows. • Led cross-vendor deployment strategy and operational readiness • Developed reusable prompt libraries and implementation standards • Improved processing time by 68% and delivered $700K+ in cost savings • Drove organization-wide adoption for vendor compliance and AI quality assurance

2026 - Present
Google Cloud Vertex AI

Team Leader, Computational Linguistics & AI Evaluation — Google (via Trilyon)

Google Cloud Vertex AIGoogle Cloud Vertex AITextText

Built and deployed Gemini-based automated agent workflows that include evaluation rubric generation and quality-control checks for enterprise vendor guidelines. Led QA/QC automation and RAG pipeline integration to improve knowledge retrieval accuracy and consistency across multilingual workflows. Enabled reusable prompt libraries and implementation standards used across large vendor operations. • Automated evaluation rubric generation for guidelines management • Performed QA/QC quality assurance to replace high-volume manual reviews • Integrated RAG pipelines with Gemini models for multilingual knowledge retrieval • Improved processing time by 68% and delivered $700K+ in cost savings through standardized QA/assessment workflows

2026 - Present

Code Annotation & AI-Generated Code Safety Evaluation — LLM Code Implementation QA

Internal/Proprietary ToolingComputer Code ProgrammingComputer Code ProgrammingComputer Programming/CodingComputer Programming/CodingFunction CallingFunction Calling

Marchand AI LLC provides specialized code annotation and AI-generated code evaluation services grounded in direct production experience assessing code outputs from frontier language models at enterprise scale. Our team evaluates not just functional accuracy but safety alignment, logical coherence, and cross-modal consistency between natural language prompts and code outputs — a distinction that separates rigorous LLM code evaluation from basic syntax checking. Google Gemini — Code Implementation Evaluation: Led bilingual evaluation of Google Gemini's code generation capabilities as part of a comprehensive multimodal assessment framework developed in collaboration with Google Cloud engineering teams. Evaluation scope included code accuracy, logical consistency, safety alignment, prompt-to-code fidelity, and agentic behavior assessment across multiple programming languages and implementation scenarios. Developed and applied scoring rubrics used directly by Google engineering teams to refine Gemini's code generation outputs across iterative model versions. Marchand AI's code annotation services include: — AI-generated code safety evaluation and content policy QA — Prompt-to-code consistency and fidelity scoring — Logical coherence and functional accuracy assessment — Adversarial prompt testing targeting code generation failure modes — Cross-modal evaluation — natural language instruction vs. code output alignment — Agentic AI code behavior assessment and boundary testing — Structured friction point reporting and RLHF-ready annotation packages — Clearance-backed discretion for proprietary code evaluation environments

2025 - Present
Labelbox

Bilingual LLM Evaluation & Linguistic Quality Annotation — English/Spanish

LabelboxLabelboxTextTextPrompt + Response Writing (SFT)Prompt + Response Writing (SFT)Evaluation/RatingEvaluation/Rating

Served as Team Lead for an 11-member computational linguistics team evaluating and instruction-tuning Google's Gemini AI models across text, voice, video, and code modalities. Designed and implemented evaluation rubrics and scoring criteria used to benchmark model performance through rigorous side-by-side (SxS) analysis. Delivered actionable reports directly informing Google Cloud engineering and product development decisions. Annotation scope included native bilingual evaluation in English and Spanish across all dialects — assessing syntactic accuracy, cultural resonance, tonal appropriateness, register consistency, and safety alignment. Developed Human-in-the-Loop quality control frameworks that increased team annotation capability by 60% over five months. Additional annotation experience includes prompt generation and safety evaluation for Meta's LLaMA AI (via Tek Systems), NLU model refinement for Toyota's in-car conversational AI (Nuance Mix), and conversational flow annotation for Poly-AI's PG&E customer service bot. Most recently, designed and delivered a full Linguistic Quality Evaluation (LQE) framework and bilingual annotation rubric for Document Tracking Services (DTS), supporting AI-assisted translation of K-12 school district compliance documentation under FERPA guidelines.

2023 - Present

Multilingual Voice AI Annotation & NLU Validation

Internal/Proprietary ToolingAudioAudioEvaluation/RatingEvaluation/RatingRed TeamingRed Teaming

Our multilingual team brings direct production experience across two major enterprise voice AI deployments. Toyota In-Car Conversational AI (Nuance Mix): Refined the NLU model for Toyota's in-car conversational AI system using Nuance Mix tooling, validating and analyzing response accuracy across multiple language inputs. Authored detailed friction point reports by probing the AI with complex, real-world user commands, identifying failure modes, dialect-specific gaps, and response inconsistencies. Utilized SQL and Python to generate structured reports and implement model adjustments directly. Poly-AI / PG&E Customer Service Bot (Google Conversational Flow Tools): Enhanced PG&E's customer service voice bot by analyzing user interactions through Google's advanced conversational flow tools. Expanded the AI's knowledge base through targeted data annotation and response correction — directly improving NLU model accuracy and user self-service capabilities for a bilingual utility customer base. Marchand AI's voice annotation services include: — Multilingual utterance collection and transcription — NLU intent and entity annotation — Conversational flow gap analysis and friction point reporting — Audio transcription, timecoded and verbatim formats — Native-level dialect validation across multiple languages upon demand

2025 - 2026