Senior AI/ML Engineer

Gesture•New York, NY

1d•Hybrid

About The Position

Gesture is where technology meets humanity -- a place where innovation, emotion, and impact collide. We're a fast-growing tech company using AI, machine learning, and intelligent logistics to power a first-of-its-kind platform that connects people and brands through real-world, tangible experiences. From our mobile app to our B2B Reach360 platform, Gesture blends data, emotion, and automation to build the future of human connection, at scale. Inside our NYC headquarters, you'll find an environment that moves with the pace and precision of Silicon Valley but with the heart of something far greater. We run on cutting-edge tools, creative experimentation, and raw ambition. Every model you build, every signal you score, every system you deploy here matters -- because it's felt and experienced by real people around the world. At Gesture, you'll work alongside some of the smartest, most driven engineers and operators in the industry -- people who think big, move fast, and care deeply about the work they do. This is a front-row seat to the future of connection. If you want to help build something that's changing how the world interacts, welcome to Gesture. WHERE WE ARE HEADED Gesture is investing aggressively in AI-driven intelligence to power the next generation of human activation and behavioral commerce. Our roadmap focuses on building intelligence that anticipates, adapts, and compounds -- predicting intent, scoring behavior in real time, and delivering deeply personalized experiences driven by real-world signal, all through systems that continuously learn and improve with every interaction. This is not surface-level automation. We are building intelligence directly into the core of how Gesture operates: how users are activated, how campaigns are scored, how brands understand and reach people at the exact moment it matters. At Gesture, you're not joining a company that bolts AI onto an existing product. You're joining a team building the Intelligence Engine -- our proprietary AI system -- from the ground up, as the core of everything we do. THE OPPORTUNITY Gesture is a late-stage, venture-backed technology company in full hyper-scale mode. We are past the experimental stage. The product works, the market is real, and the Intelligence Engine is what powers our differentiation. What we need now is an engineer who can build, ship, and own AI and ML systems that perform in production -- not in a notebook, not in a pilot, and not behind a research wall. Someone who treats model quality the same way a revenue operator treats a missed number: personally. The Senior AI/ML & Data Engineer Senior AI/ML Engineer is the technical backbone of the Intelligence Engine. You sit at the intersection of data science, machine learning, and software engineering -- designing the systems that score behavior, surface insight, and activate outcomes across every surface of the Gesture platform. Let's be direct about the environment: this is not a place for researchers who hand off to "implementation teams." It is not a place for engineers who prototype endlessly without shipping, who treat model accuracy as someone else's problem, or who need perfect data to start building. There are no layers to hide behind, no slow approval chains, and no tolerance for the gap between what models do in testing and what they do in production. If your instinct when a model underperforms is to document it in a ticket, stop reading here. If your instinct is to fix it, keep going. WHAT THIS ROLE ACTUALLY IS You will personally own: The Intelligence Engine -- Scoring & Activation -- The behavioral scoring algorithms that power user-facing activation across B2C and B2B; you own the logic, the lift, and the outcomes ML Pipeline Architecture -- The full lifecycle: data ingestion, feature stores, model training, evaluation, deployment, monitoring, and retraining triggers LLM Integration & Optimization -- Integration, fine-tuning, and production deployment of large language models for contextual inference, personalization, and behavioral pattern recognition Behavioral Signal Processing -- Extraction of meaningful features from structured and unstructured behavioral data; you define what signals matter and why Model Evaluation & Governance -- Evaluation frameworks, A/B testing infrastructure, bias detection, and drift monitoring; model integrity in production is non-negotiable Data Infrastructure -- Scalable pipeline and schema design that supports real-time and batch inference, built in collaboration with engineering Research to Production -- The full path from experimentation to deployed, monitored, and maintained production systems; nothing lives in a notebook forever Cross-Functional Technical Partnership -- Direct collaboration with product, engineering, and leadership to translate business outcomes into machine learning problem definitions You will own the intelligence layer of the product. You will not "support" it. WHAT THIS ROLE IS NOT Let's be clear: This is not a research role with no path to production This is not a data analyst position This is not a "build models and hand off to engineering" job This is not a role for engineers who treat deployment, monitoring, and retraining as someone else's problem This is not a role for people who need clean data, perfect infrastructure, or a fully staffed ML team to be effective If you have never owned a model in production -- including when it breaks, drifts, or underperforms -- this will be painful WHAT YOU WILL BE HELD ACCOUNTABLE FOR You own: Model performance in production, not just in evaluation; lift, precision, recall, and business impact are your metrics End-to-end ML pipeline reliability: data in, scored output out, on time, every time Speed from experimentation to production; the faster a good idea gets to users, the better Signal quality and feature engineering that makes models better than the data suggests they should be Model governance: version control, drift detection, bias auditing, and documentation that holds up to scrutiny Infrastructure decisions that scale without requiring a rewrite every six months Cross-functional clarity: engineering, product, and leadership should always understand what the models do, why they do it, and how to measure whether it's working The technical roadmap for the Intelligence Engine; where we are, where we're going, and what's blocking progress When the Intelligence Engine performs and the product is smarter because of your work, you get credit. When model quality degrades and no one caught it, that falls on you. WHAT YOU WILL OWN I. The Intelligence Engine -- Scoring & Behavioral Intelligence Design, build, and continuously improve the behavioral scoring system that powers Gesture's core activation model across B2C and B2B Own signal selection, feature engineering, and model architecture decisions for all scoring use cases Define and enforce evaluation standards for every model that touches a user-facing surface Build the feedback loops that keep models improving as behavior data compounds II. ML Pipeline & Infrastructure Own the full ML lifecycle from raw data to deployed inference: ingestion, transformation, training, evaluation, deployment, and monitoring This includes owning the raw data layer - you are not handed clean features. You build and maintain ETL/ELT pipelines from third-party data sources (APIs, webhooks, flat files), handle auth, rate limits, and schema drift, and design event-driven ingestion using Pub/Sub and Dataflow for real-time signal processing. Design pipeline architecture for both real-time and batch inference use cases Build and maintain the feature store, experiment tracking, and model registry that make the ML organization productive at scale Partner with engineering to ensure ML infrastructure is production-grade, not prototype-grade III. LLM Integration & Applied AI Integrate and optimize large language models for contextual inference, personalization, and behavioral pattern detection across Gesture's platform Evaluate and implement retrieval-augmented generation, fine-tuning, and embedding workflows based on product requirements Make rigorous tradeoff decisions on model selection: cost, latency, accuracy, and reliability in production Build and maintain vector search and semantic retrieval capabilities that power personalization at scale IV. Data Science & Experimentation Design and execute experiments that produce statistically valid, business-relevant results Own A/B testing infrastructure and experiment design standards across the ML team Develop causal inference and multi-armed bandit frameworks where appropriate to accelerate learning velocity Translate experiment results into model improvements, product decisions, and actionable recommendations for leadership V. Model Governance & Quality Build and maintain model monitoring dashboards that surface degradation, drift, and anomalies before they become product issues Own bias auditing and fairness evaluation across all scoring and classification systems Maintain documentation, versioning, and rollback capability for all production models Establish and enforce model quality standards across the team as the ML function scales VI. Cross-Functional Technical Leadership Translate business problems into machine learning problem definitions with clear success criteria Communicate model behavior, performance, and limitations to non-technical stakeholders without losing precision Partner with product and engineering to sequence ML work against business priorities Build the technical foundation that allows a growing ML team to operate without creating chaos

Requirements

You have 5+ years of hands-on ML engineering or applied AI experience with models running in production, not just in research or evaluation
You have personally owned the full ML lifecycle: data, features, training, deployment, monitoring, and retraining
You are proficient in Python,SQL and have production experience with ML frameworks (PyTorch, TensorFlow, scikit-learn, XGBoost, or equivalent)
You have designed and deployed LLM-powered features: RAG pipelines, fine-tuning workflows, vector embeddings, or semantic search in production
You understand data modeling, feature engineering, and pipeline design for both real-time and batch inference at scale
You have experience with ML orchestration tools (Airflow, Prefect, Kubeflow, or similar)
You have built and maintained production data pipelines from external sources - not just consumed them; you've handled API integration failures, schema drift, and data quality issues upstream of any model
You have experience with event-driven pipeline architecture: Pub/Sub, Dataflow, or equivalent
You write clean, production-grade code -- not prototype code that needs to be rewritten before it ships
You are comfortable with cloud-native ML infrastructure (GCP preferred: Vertex AI, BigQuery, Cloud Run)
Designing and building scalable systems for model development, training, and deployment, managing large-scale data pipelines and distributed compute environments.
You can make model architecture decisions under ambiguity and course-correct fast when the data tells you to
You communicate technical concepts to non-technical stakeholders without losing accuracy

Nice To Haves

Experience building behavioral analytics, engagement scoring, or personalization engines in a consumer or B2B product context
Familiarity with vector databases (Pinecone, Weaviate, pgvector) and embedding-based retrieval workflows
Background in experiment design, multi-armed bandits, or causal inference for production decision systems
Exposure to mobile-first AI systems or consumer product intelligence at scale
Comfort operating lean -- resourceful and inventive when the infrastructure isn't perfect and the playbook doesn't exist yet
Experience scaling an ML function and mentoring engineers as the team grows

Responsibilities

Design, build, and continuously improve the behavioral scoring system that powers Gesture's core activation model across B2C and B2B
Own signal selection, feature engineering, and model architecture decisions for all scoring use cases
Define and enforce evaluation standards for every model that touches a user-facing surface
Build the feedback loops that keep models improving as behavior data compounds
Own the full ML lifecycle from raw data to deployed inference: ingestion, transformation, training, evaluation, deployment, and monitoring
Design pipeline architecture for both real-time and batch inference use cases
Build and maintain the feature store, experiment tracking, and model registry that make the ML organization productive at scale
Partner with engineering to ensure ML infrastructure is production-grade, not prototype-grade
Integrate and optimize large language models for contextual inference, personalization, and behavioral pattern detection across Gesture's platform
Evaluate and implement retrieval-augmented generation, fine-tuning, and embedding workflows based on product requirements
Make rigorous tradeoff decisions on model selection: cost, latency, accuracy, and reliability in production
Build and maintain vector search and semantic retrieval capabilities that power personalization at scale
Design and execute experiments that produce statistically valid, business-relevant results
Own A/B testing infrastructure and experiment design standards across the ML team
Develop causal inference and multi-armed bandit frameworks where appropriate to accelerate learning velocity
Translate experiment results into model improvements, product decisions, and actionable recommendations for leadership
Build and maintain model monitoring dashboards that surface degradation, drift, and anomalies before they become product issues
Own bias auditing and fairness evaluation across all scoring and classification systems
Maintain documentation, versioning, and rollback capability for all production models
Establish and enforce model quality standards across the team as the ML function scales
Translate business problems into machine learning problem definitions with clear success criteria
Communicate model behavior, performance, and limitations to non-technical stakeholders without losing precision
Partner with product and engineering to sequence ML work against business priorities
Build the technical foundation that allows a growing ML team to operate without creating chaos

Benefits

Full benefits: health, dental, vision, equity and stock options
Make a Gesture Day: Each employee receives the chance to earn credit to send a surprise gift to someone they admire -- living the company mission firsthand
Gesture Swag Pack: Exclusive branded gear (hoodie, water bottle, or cap)
Beautiful HQ Experience: Modern New York office with creative work zones, media studio, and collaboration spaces
Paid Team Lunches or Coffee Hours: Casual networking with leadership and cross-department peers
Professional development budget

Stand Out From the Crowd

Upload your resume and get instant feedback on how well it matches this job.

Upload and Match Resume