Senior Machine Learning Engineer, Conversational AI

at Dropbox

Remote

Role Description

Machine Learning Engineers at Dropbox develop high-impact products that advance Dropbox’s mission of designing a more enlightened way of working. Dropbox MLE’s deliver products using the full range of ML techniques—including computer vision, natural language processing, recommendation systems, large language models (LLMs), and LLM-driven agents—and modern practices such as LLM fine-tuning, prompt engineering, and retrieval-augmented generation (RAG).While some of our algorithms run on mobile devices, others require large clusters on our infrastructure. Accordingly, we apply techniques like model compression, distillation, on-device optimization, and scalable training/serving with robust evaluation and monitoring.

Dropbox is looking for Machine Learning Engineers with an academic or practical background in ML, ideally with experience in natural language understanding, information retrieval, knowledge extraction, or deep learning, as well as hands-on work with LLMs (pre-training and fine-tuning, including instruction tuning and LoRA/adapters), prompt engineering, RAG pipelines, safety/guardrails and evaluation, and productionizing generative and predictive systems.

Our Engineering Career Framework is viewable by anyone outside the company and describes what’s expected for our engineers at each of our career levels. Check out our blog post on this topic and more here.

Responsibilities

Design, code, train, test, deploy, and iterate on large-scale ML and LLM systems across cloud and mobile/edge environments
Build delightful, privacy-first product experiences (e.g., intelligent search, document understanding, recommendations, and AI assistants) in partnership with Engineering, Product, and Design
Lead end-to-end LLM workflows: data curation, prompt engineering, retrieval-augmented generation (RAG)pipelines, tool use/agents, and fine-tuning (e.g., instruction tuning, LoRA/adapters) with rigorous evaluation
Develop and maintain production-quality services for training and serving, including scalable APIs, vector/feature stores, and streaming/ETL data pipelines
Optimize for latency, cost, and quality using techniques like quantization, distillation, caching, batching, and autoscaling; tailor models for on-device vs. cluster execution
Establish robust offline/online evaluation: experiment design, A/B testing, guardrails and safety checks, hallucination mitigation, and automated monitoring/observability with clear SLOs
Communicate technical trade-offs, risks, and impact to cross-functional stakeholders; write clear design docs, roadmaps, and decision records
Partner with Security, Legal, and Privacy to ensure responsible AI, data governance, and compliance in training and inference
Proactively explore and integrate advances in ML/AI (CV, NLP, recsys, LLMs) and rapidly prototype and transfer promising research into production
Mentor teammates, contribute to code reviews and best practices, and help shape the technical direction of ML and AI at Dropbox

Many teams at Dropbox run Services with on-call rotations, which entails being available for calls during both core and non-core business hours. If a team has an on-call rotation, all engineers on the team are expected to participate in the rotation as part of their employment. Applicants are encouraged to ask for more details of the rotations to which the applicant is applying.

Requirements

BS or MS in Computer Science or related technical field involving Machine Learning or equivalent technical experience
8+ years of experience in engineering with 5+ years of experience building Machine Learning or AI systems
Strong industry experience working with large scale data
Strong analytical and problem-solving skills
Familiarity with search-related applications of Large Language Models
Proven software engineering skills across multiple languages including but not limited to Python, Go, C/C++
Experience with Machine Learning software tools and libraries (e.g., PyTorch, HuggingFace, TensorFlow, Keras, Scikit-learn, etc.)

Preferred Qualifications

PhD in Computer Science or related field with research in machine learning
Experience with one or more of the following: natural language processing, deep learning, bayesian reasoning, recommender systems, learning to rank, speech processing, learning from semistructured data, graph learning, reinforcement or active learning, large language models, ML software systems, retrieval-augmented generation, machine learning on edge devices
Experience building 0→1 ML products at large (dropbox-level) scale or multiple 0→1 products at smaller scale including experience with large-scale product systems

Compensation

US Zone 1

$236,300—$319,700 USD

US Zone 2

$212,700—$287,700 USD

US Zone 3

$189,000—$255,800 USD