Staff AI Engineer/AI Architect, Conversation Intelligence Systems (On-Site, New York)

Xenoss; Northern, New York (KY, NY); Hybrid; USD 150,000 - 190,000; Full time

Job summary

Xenoss in New York is seeking a Staff AI Engineer/AI Architect to lead the architecture of a real-time conversation intelligence system for a major financial services client. You will guide data taxonomy, model training, evaluation, and production readiness across the In-Call Assistant lifecycle.

You will own end-to-end design, collaborate with AI engineers, data teams, MLOps, and client SMEs, and balance accuracy, latency, and governance while delivering PoC-to-production solutions with strong

Qualifications

• Must have hands-on production AI/ML experience.
• Experience with NLP, conversational AI, or transcript-based intelligence systems.
• Ability to design evaluation frameworks, not just run experiments.
• Experience building or validating structured datasets from unstructured text.
• Strong understanding of LLM-based extraction, classification, RAG, and fine-tuning trade-offs.
• Practical knowledge of classical ML or predictive modeling.
• Understanding of probability-based prediction, calibration, and outcome evaluation.
• Comfort working with messy enterprise data and incomplete labels.
• Ability to communicate with both technical teams and business stakeholders.
• Strong ownership of ambiguity, scope control, and PoC validation strategy.

Responsibilities

• Design end-to-end AI architecture for the In-Call Assistant.
• Define signal and trigger taxonomies for live conversations.
• Design training strategies for signal detection and specialist recommendation models.
• Shaping data preparation, annotation, and SME validation workflows.
• Evaluating fine-tuning, post-training, RAG, and hybrid approaches.
• Designing low-latency signal detection, routing, context preparation, and confidence management.
• Defining grounding, guardrails, abstention, and policy-compliance behavior.
• Making trade-offs between model quality, latency, cost, explainability, and governance.
• Partnering with AI engineers, data engineering, MLOps, and client SMEs.

Skills

Applied AI/ML systems

NLP/Conversational AI

Evaluation frameworks

Structured datasets from unstructured

LLM extraction & classification

Model governance

Low-latency inference

Stakeholder communication

Ambiguity management

Tools

PyTorch

Hugging Face

Job description

Staff AI Engineer/AI Architect, Conversation Intelligence Systems (On-Site, New York)

Who weare

Xenoss isanAI engineering and integration services company, helping medium tolarge enterprises runAI transformation end-to-end, from situation analysis and goals framing todata discovery and preparation, pipeline building, model development, retraining pipeline design, solution deployment, and support.

Webuild abroad spectrum ofAI solutions such asuser behaviour prediction, content generation, NLP, audience segmentation, pathfinding solutions, AIassistants, edge computer vision, fraud detection, and others.

Wework with prominent companies such asMicrosoft, Toshiba, AstraZeneca, Activision Blizzard, Verve Group, Voodoo Games, and Telefonica, among others.

We’re included inthe top 100 software companies onthe Inc.5000list.

What isthe project

We’re hiring aStaff AIEngineer/ AISolution Architect tolead theAI architecture ofalong-term In-Call Assistant initiative for aworld-leading financial services company.

The project focuses onbuilding areal-time conversationalAI system that supports front-office employees during live customer conversations. The system identifies customer needs, objections, buying signals, and required process steps, and provides concise, context-aware recommendations.

The solution combines low-latency signal detection, context preparation, specialist recommendation generation, RAG over approved product and policy knowledge, confidence management, and compliance guardrails.

You will help define how the AIarchitecture, models, evaluation framework, and feedback loops are designed and evolved from the initial offline version tolive production use

What will you do

You’ll lead the appliedAI architecture across the In-Call Assistant lifecycle, from data and taxonomy design tomodel training, evaluation, and production readiness.

Core work includes:

• Designing the end-to-end AIarchitecture for the In-Call Assistant
• Defining signal and trigger taxonomies for live conversations
• Designing training strategies for signal detection and specialist recommendation models
• Shaping data preparation, annotation, and SME validation workflows
• Evaluating fine-tuning, post-training, RAG, and hybrid approaches
• Designing low-latency signal detection, routing, context preparation, and confidence management
• Designing evaluation frameworks, golden datasets, and model improvement cycles
• Defining grounding, guardrails, abstention, and policy-compliance behavior
• Making trade-offs between model quality, latency, cost, explainability, and governance
• Partnering with AIengineers, data engineering, MLOps, and client SMEs

Technology landscape

You’ll operate across the modern appliedAI and MLecosystem, including:

• LLM and smaller-model training for conversational AI
• SFT, DPO/ preference optimization, LoRA/ QLoRA, and PEFT
• PyTorch and Hugging Face ecosystem
• Signal extraction and multi-label classification
• RAG and knowledge-grounded recommendation generation
• Embeddings, retrieval, and context preparation
• Low-latency model serving and inference optimization
• Golden dataset creation, annotation, and SME validation
• Model evaluation, confidence calibration, and error analysis
• MLOps, monitoring, feedback loops, and model governance

Scope ofownership and delivery context

AtStaff/Architect level, you’ll own the appliedAI architecture and evaluation strategy for acomplex enterprise AIprogram.

Core ownership

• Define theAI approach for the conversation intelligence PoC
• Establish the event/ intent/ insight taxonomy
• Define the golden dataset strategy and annotation workflow
• Establish evaluation frameworks and acceptance criteria
• Drive trade-offs between accuracy, explainability, latency, cost, and governance
• Decide which modeling approaches are appropriate for each use case
• Act asanescalation point for AIarchitecture, evaluation, and data strategy decisions

Team and delivery context

• Coordination and collaboration with the client team
• Work within across-functional team spanning AIengineering, data engineering, MLOps, solution architecture, and client stakeholders
• Partner with domain SMEs ontaxonomy, labeling, and validation
• Mentor engineers working onextraction, evaluation, and data pipelines
• Translate ambiguous business use cases into testableAI hypotheses and validation plans

What should you bring

Must have

• Strong hands-on experience with applied AI/ MLsystems inproduction-oriented environments
• Experience with NLP, conversationalAI, ortranscript-based intelligence systems
• Ability todesign evaluation frameworks, not just run experiments
• Experience building orvalidating structured datasets from unstructured text
• Strong understanding ofLLM-basedextraction, classification, RAG, and fine-tuning trade-offs
• Practical knowledge ofclassical MLor predictive modeling
• Understanding ofprobability-based prediction, calibration, and outcome evaluation
• Comfort working with messy enterprise data and incomplete labels
• Ability tocommunicate with both technical teams and business stakeholders
• Strong ownership ofambiguity, scope control, and PoC validation strategy

Nice tohave

• Financial services domain exposure
• Experience with sales, call center, orcustomer conversation analytics
• Speech/ ASR pipeline familiarity
• Model governance and auditability experience
• Experience with real-time AIsystems orlow-latency inference
• Experience combining unstructured conversation signals with structured CRM, transaction, orcustomer profile data
• Experience designing golden datasets and SME review workflows

Operating model

• Engagement structure: FTE-equivalent via long-term B2B contract
• Work location: On-site orclosely aligned with the client team inNew York
• Infrastructure: Client environment only, noexternal training ordata processing environments
• Data residency: All work executed within the client perimeter
• Delivery mode: PoC-first, with apath toward production-grade conversation intelligence and prediction systems