Proximity Works

At Proximity Works, we build AI-native products that operate at scale. We're now entering a critical phase—moving beyond baseline functionality to delivering high-quality, delightful user experiences across search, conversational AI, and multimodal systems.

This role sits at the intersection of LLMs, search, personalization, and video intelligence, with direct impact on products used by hundreds of millions of users.

What You'll Work On

• Improving conversational search quality—making responses not just correct, but intuitive, contextual, and delightful
• Contributing to the unification of traditional and conversational search, including routing logic, ranking systems, and cost optimization
• Building and integrating ranking and personalization systems, combining user context with query intent
• Developing video intelligence capabilities—scene-level understanding, embedding-based retrieval (CLIP and similar), and dynamic content stitching
• Working on reasoning over unstructured data (RAG, multi-hop reasoning), especially for high-context domains like sports and content discovery
• Designing and iterating on evaluation frameworks to continuously improve quality and performance
• Operating in fast iteration cycles, closely aligned with product and engineering teams

Key Problem Areas

• Elevating search experience from functional to "magical"
• Designing a unified search architecture across modalities and interfaces
• Balancing quality, latency, and cost in production systems
• Scaling systems while maintaining high standards (the "90 → 95" problem)

Must-Have

• Hands-on experience building and shipping LLM/GenAI-powered products (beyond standalone model training)
• Strong foundation in search, NLP, or information retrieval (ranking, query understanding, retrieval systems)
• Experience designing evaluation pipelines for generative AI (quality tracking, regression detection, metric-driven iteration)
• Ability to thrive in fast-paced, iterative environments with evolving product requirements

Good to Have

• Experience with multimodal systems (video/image embeddings, CLIP, temporal video understanding)
• Background in personalization or recommendation systems, especially within search
• Experience working on quality optimization at scale
• Strong intuition for cost-efficient system design (latency, inference cost, routing strategies)

Requirements

Must-Have

• Experience: 4.5-10 years in applied Data Science/ML with a strong focus on generative AI.
• Hands-on experience building and shipping LLM/GenAI-powered products (beyond standalone model training)
• Strong foundation in search, NLP, or information retrieval (ranking, query understanding, retrieval systems)
• Experience designing evaluation pipelines for generative AI (quality tracking, regression detection, metric-driven iteration)
• Ability to thrive in fast-paced, iterative environments with evolving product requirements

Good to Have

• Experience with multimodal systems (video/image embeddings, CLIP, temporal video understanding)
• Background in personalization or recommendation systems, especially within search
• Experience working on quality optimization at scale
• Strong intuition for cost-efficient system design (latency, inference cost, routing strategies)

Benefits

• Best in class salary: We hire only the best, and we pay accordingly.
• Proximity Talks: Meet other designers, engineers, and product geeks — and learn from experts in the field.
• Keep on learning with a world-class team: Work with the best in the field, challenge yourself constantly, and learn something new every day.

About us

Proximity is the trusted technology, design, and consulting partner for some of the biggest Sports, Media, and Entertainment companies in the world! We’re headquartered in San Francisco and have offices in Palo Alto, Dubai, Mumbai, and Bangalore. Since 2019, Proximity has created and grown high-impact, scalable products used by 370 million daily users, with a total net worth of $45.7 billion among our client companies.

Today, we are a global team of coders, designers, product managers, geeks, and experts. We solve complex problems and build cutting-edge tech, at scale. Our team of Proxonauts is growing quickly, which means your impact on the company’s success will be huge. You’ll have the chance to work with experienced leaders who have built and led multiple tech, product, and design teams. 

About the Role

We're looking for a hands-on ML Engineer who combines strong machine learning fundamentals with the backend infrastructure instincts needed to take models from research into reliable, scalable production systems. This role sits at the intersection of cutting-edge generative video AI and the MLOps discipline required to run it at scale.

You'll work closely with backend, platform, and content teams to deliver high-performance ML components under strict quality, latency, and throughput requirements.

Key Responsibilities

• Train, fine-tune, and evaluate generative video and multimodal models (image-to-video, text-to-video, lip-sync, character consistency)
• Build and manage end-to-end ML pipelines: data ingestion, preprocessing, training, evaluation, and versioning
• Own model deployment and serving infrastructure — containerization, GPU-optimized inference, model registries, and rollout strategies
• Implement MLOps best practices: experiment tracking, model monitoring, drift detection, A/B testing, and observability
• Design and maintain scalable inference systems optimized for low latency, high throughput, and cost-efficient GPU utilization
• Develop caching and batching strategies to meet SLA targets in production video generation workflows
• Collaborate with backend engineering teams on integrating ML services into distributed systems
• Contribute to long-term roadmap: foundational model training strategies, LoRA fine-tuning pipelines, and multi-character generalization

Requirements

Required Qualifications

• 4-10 years of experience in Machine Learning / Applied ML Engineering
• Strong fundamentals in deep learning, Transformers, and generative model architectures
• Hands-on experience with model training at scale — including fine-tuning large models (LoRA, full fine-tune) on custom datasets
• Solid MLOps experience: experiment tracking (MLflow, W&B), CI/CD for ML, model versioning, and serving frameworks (Triton, TorchServe, vLLM, or equivalent)
• Strong Python skills and fluency with PyTorch and the modern ML stack
• Experience deploying and operating ML systems in distributed cloud environments (GCP, AWS, or Azure) — GPU provisioning, autoscaling, and cost management
• Comfort working on ambiguous, high-impact problems with cross-functional teams

Preferred Qualifications

• Experience with video generation, diffusion models, or multimodal architectures (DiT, U-Net, audio-video joint models)
• Familiarity with LoRA/IC-LoRA fine-tuning workflows for character or identity consistency
• Experience in media, OTT, sports, or large-scale content platforms
• Knowledge of inference optimization techniques: quantization (FP8/INT8), batching, async orchestration, and GPU memory management
• Exposure to TTS/voice cloning systems or audio-video synchronization pipelines

Benefits

What you get

• Best in class salary: We hire only the best, and we pay accordingly.
• Proximity Talks: Meet other designers, engineers, and product geeks — and learn from experts in the field.
• Keep on learning with a world-class team: Work with the best in the field, challenge yourself constantly, and learn something new every day.

About us

Proximity is the trusted technology, design, and consulting partner for some of the biggest Sports, Media, and Entertainment companies in the world! We’re headquartered in San Francisco and have offices in Palo Alto, Dubai, Mumbai, and Bangalore. Since 2019, Proximity has created and grown high-impact, scalable products used by 370 million daily users, with a total net worth of $45.7 billion among our client companies.

Today, we are a global team of coders, designers, product managers, geeks, and experts. We solve complex problems and build cutting-edge tech, at scale. Our team of Proxonauts is growing quickly, which means your impact on the company’s success will be huge. You’ll have the chance to work with experienced leaders who have built and led multiple tech, product, and design teams.