How to Apply to Pony.ai

10 min read Last updated March 7, 2026 4 open positions
Blake Crosley
Blake Crosley Researched using publicly available information, hiring data, and ATS analysis.

Key Takeaways

  • Tailor your resume for every Pony.ai role you apply to — mirror the exact technical keywords from the job description (e.g., 'TensorRT,' 'perception infrastructure,' 'model quantization') to maximize visibility in Workable's keyword search and filtering.
  • Build a visible publication or portfolio trail: link your Google Scholar, arXiv preprints, and GitHub repositories directly in your resume header and in Workable's optional URL fields — Pony.ai's roles explicitly value research output.
  • Prepare for system design questions specific to autonomous driving infrastructure — practice designing real-time perception pipelines, model serving architectures, and data annotation systems at scale before your interview.
  • Study Pony.ai's published technical work, blog posts, and any publicly shared research to demonstrate genuine domain interest during interviews — interviewers reward candidates who've done their homework on the company's specific technical approach.
  • Use a clean, single-column PDF resume with standard section headers to ensure Workable's parser accurately extracts your qualifications — a parsing error on your degree or skills could cost you the screening round.
  • For internship roles, clearly state your degree level, program, expected graduation date, and semester availability in your education section — Pony.ai's intern postings are semester-specific and degree-gated, so ambiguity slows your candidacy.

About Pony.ai

Pony.ai is a global autonomous driving technology company on a mission to revolutionize transportation by delivering safe, reliable self-driving vehicles at scale. Founded in 2016 by James Peng and Tiancheng Lou — both former leaders at Baidu's autonomous driving division — the company has rapidly established itself as one of the most technically ambitious players in the Level 4 autonomous driving space. Pony.ai went public on NASDAQ in late 2024, a milestone that underscored investor confidence in its path to commercialization. The company operates across the United States and China, with R&D centers in Silicon Valley (Fremont, CA), Beijing, Shanghai, and Guangzhou. Its technology stack spans the full autonomous driving pipeline: perception, prediction, planning, and vehicle control — all powered by deep learning and advanced sensor fusion across cameras, LiDAR, and radar. Pony.ai has deployed robotaxi and robotruck services in multiple cities and maintains a strategic partnership with Toyota. Culturally, Pony.ai attracts engineers and researchers who thrive at the intersection of cutting-edge machine learning and real-world robotics. Teams are lean, technically rigorous, and move fast — a hallmark of its startup DNA even as the company scales post-IPO. The environment rewards intellectual curiosity, deep technical ownership, and a willingness to tackle problems where software meets the physical world. For engineers who want their code to literally drive on public roads, Pony.ai offers a rare combination of research depth and tangible, high-stakes impact.

Application Process

  1. 1
    Identify Your Target Role on Pony.ai's Workable Portal

    Visit Pony.ai's careers page at apply.workable.com/pony-dot-ai/ and review the active openings, which are heavily concentrated in machine learning, perception, and deep learning engineering. Read each job description carefully — Pony.ai's postings tend to be technically specific, listing exact framework requirements (e.g., PyTorch, TensorRT), publication expectations, and degree-level preferences. Note whether the role is a full-time position or an internship, as intern roles often specify semester windows and degree requirements (Master's or PhD).

  2. 2
    Prepare a Technically Dense, AV-Relevant Resume

    Before clicking 'Apply,' tailor your resume to reflect autonomous driving and robotics-adjacent experience. Pony.ai's hiring managers are domain experts, so surface relevant work in perception systems, sensor fusion, deep learning model optimization, real-time inference, or large-scale data pipelines. Quantify your contributions — inference latency reductions, model accuracy improvements, or dataset scale — because these are the metrics that matter in AV engineering.

  3. 3
    Submit Your Application Through Workable

    Complete the Workable application form, which typically asks for your resume, contact information, LinkedIn profile, and sometimes a cover letter or additional links (such as a GitHub or Google Scholar profile). Workable's interface is streamlined, but don't rush — fill in every optional field, as completeness signals genuine interest. If you have relevant publications or open-source contributions, include direct links rather than assuming the reviewer will search for them.

  4. 4
    Initial Resume and Application Screening

    Pony.ai's recruiting team, aided by Workable's parsing and filtering tools, will review your application against the role's technical requirements. Given the company's lean team structure and highly specialized roles, this screen is typically rigorous — expect close scrutiny of your degree background, specific framework proficiencies, and any autonomous driving or robotics experience. Candidates with published research in relevant venues (CVPR, NeurIPS, ICRA, ICLR) or demonstrable experience with real-time ML systems commonly advance.

  5. 5
    Technical Phone Screen or Recruiter Conversation

    Candidates who pass the initial screen are typically contacted for a phone or video interview. For engineering roles at AV companies like Pony.ai, this often includes a coding component or a deep technical discussion about your past projects. Be prepared to explain your contributions to specific systems — not just what a team built, but what you personally designed, debugged, and shipped.

  6. 6
    On-Site or Virtual Technical Interview Loop

    The core interview stage at Pony.ai commonly involves multiple rounds: algorithm and data structure coding sessions, system design focused on autonomous driving infrastructure (e.g., designing a perception pipeline, optimizing inference for edge deployment), and deep dives into your ML research or engineering portfolio. Some rounds may involve discussing specific papers or presenting your published work. Expect interviewers who are senior engineers or research scientists — these are peer-level technical conversations, not HR screens.

  7. 7
    Offer, Negotiation, and Onboarding

    Successful candidates receive an offer that typically includes competitive compensation, equity (particularly relevant post-IPO), and benefits. Given Pony.ai's dual US-China operations, onboarding logistics may vary by office location. Respond promptly to any offer communication and ask clarifying questions about team placement, project scope, and growth trajectory — these details matter at a company where individual engineers carry significant ownership.

Resume Tips for Pony.ai

critical

Lead with Autonomous Driving and Robotics Keywords

Pony.ai's roles center on perception, sensor fusion, deep learning inference, and real-time systems. Your resume must explicitly include terms like 'LiDAR point cloud processing,' 'camera-based 3D object detection,' 'model quantization,' 'TensorRT,' 'ONNX,' 'real-time inference,' or 'on-vehicle deployment' if they're truthfully in your background. Workable's search and filtering tools allow recruiters to surface candidates by keyword, so burying relevant skills in dense paragraphs risks being overlooked. Place these terms in your skills section, job titles, and project descriptions.

critical

Quantify ML and Systems Performance Improvements

AV engineering is a field where milliseconds and percentage points have life-or-death implications. Instead of writing 'improved model performance,' state 'reduced inference latency by 40% on embedded GPU through TensorRT optimization while maintaining mAP above 78%.' Pony.ai's engineering culture values precision — your resume should reflect the same rigor. Include dataset sizes, training time reductions, accuracy benchmarks, and deployment environments wherever possible.

critical

Highlight Publications and Research Contributions

Multiple Pony.ai roles — especially Research Intern and MLE positions — explicitly value published work. If you have papers at top-tier venues (CVPR, NeurIPS, ECCV, ICRA, CoRL, ICLR), list them prominently with citation counts if notable. Even workshop papers or preprints on arXiv demonstrate research engagement. Link to your Google Scholar profile in your contact header so reviewers can quickly assess your publication record.

recommended

Specify Framework and Language Proficiency Precisely

Pony.ai's job descriptions call out specific tools: PyTorch, C++, CUDA, TensorRT, and Python are recurring requirements. Don't list generic 'programming skills' — instead, specify 'C++17 (production-level, 4 years)' or 'PyTorch (custom operators, distributed training).' This level of specificity helps Workable's keyword matching and immediately signals to a technical reviewer that you're operating at the right depth. If you've contributed to open-source ML frameworks, mention the project by name.

recommended

Use a Clean, Single-Column Format for Workable Parsing

Workable's resume parser handles standard formats well but can struggle with multi-column layouts, tables, headers/footers, and heavily designed templates. Use a single-column layout with clearly labeled sections (Education, Experience, Publications, Skills). Submit as PDF unless the posting specifies otherwise. Avoid embedding important information in images, charts, or text boxes — the parser may skip them entirely, meaning your key qualifications never reach the recruiter's screen.

recommended

Demonstrate Full-Stack ML Ownership, Not Just Modeling

Pony.ai's infrastructure and optimization roles require engineers who work across the entire ML lifecycle — from data pipeline construction and model training to deployment and on-vehicle runtime optimization. Show that you've owned end-to-end projects, not just trained models in notebooks. Mention experience with CI/CD for ML, model serving, A/B testing in production, or hardware-aware neural architecture design. This distinguishes you from purely academic applicants.

nice_to_have

Include Degree Details and Expected Graduation for Intern Roles

Pony.ai's internship postings specify degree level (Master's or PhD) and semester availability (e.g., Spring 2026). Make your education section unambiguous: include your degree type, program name, university, expected graduation date, and advisor name if relevant to ML/robotics research. Failing to include these details can slow down your screening, as recruiters need to quickly verify eligibility for semester-specific intern cohorts.

nice_to_have

Add a GitHub or Portfolio Link with Relevant Repositories

AV and ML roles benefit enormously from tangible code samples. If you have public repositories demonstrating work in computer vision, point cloud processing, model optimization, or robotics simulation, link them directly in your resume header. Pin your most relevant repos and ensure they have clear README files. Pony.ai's technical reviewers may check your code quality, commit history, and documentation habits as a proxy for engineering maturity.

ATS System: Workable

Workable is a cloud-based applicant tracking system widely used by growth-stage technology companies like Pony.ai. It parses uploaded resumes to extract structured data — contact info, work history, education, and skills — then allows recruiters to search, filter, and rank candidates by keyword and qualification match. Understanding how Workable processes your application gives you a structural advantage in getting surfaced to Pony.ai's hiring team.

  • Submit your resume as a single-column PDF — Workable's parser handles this format most reliably and avoids the text extraction errors common with .docx files that use complex formatting.
  • Place critical keywords (PyTorch, C++, CUDA, perception, LiDAR, deep learning, TensorRT) in both your skills section and within job description bullet points, since Workable indexes both for search filtering.
  • Avoid using headers, footers, or text boxes for important information — Workable's parser frequently skips content in these elements, potentially dropping your degree or contact details.
  • Use standard section headings like 'Experience,' 'Education,' 'Skills,' and 'Publications' — Workable maps content to its structured fields based on these conventional labels.
  • Fill out all optional fields in the Workable application form (LinkedIn URL, portfolio, additional links) — recruiters at lean companies like Pony.ai often use these to quickly assess candidate depth without scheduling a call.
  • If applying to multiple Pony.ai roles, tailor your resume for each submission — Workable tracks applications per position, and a generic resume that doesn't match a role's specific keywords may be filtered out at the search stage.

Complete Workable Resume Guide →

Interview Culture

Pony.ai's interview process reflects its identity as a deep-tech autonomous driving company staffed by researchers and engineers from top AI labs and universities.

Expect a rigorous, technically dense evaluation that prioritizes depth of knowledge over breadth. The process typically begins with a recruiter screen or technical phone interview lasting 30-60 minutes. For software engineering and MLE roles, this first round commonly includes a live coding exercise focused on algorithms, data structures, or applied ML problem-solving — think LeetCode-medium-to-hard difficulty, often with a twist toward practical systems or geometry problems relevant to autonomous driving (spatial reasoning, coordinate transforms, efficient nearest-neighbor search in point clouds). The core interview loop — conducted on-site at Pony.ai's Fremont office or virtually — generally involves 3-5 sessions across a single day or split over multiple days. These sessions typically cover: (1) algorithmic coding rounds with an emphasis on C++ and Python, (2) system design focused on real-time ML infrastructure — you might be asked to design a perception pipeline, a model serving system, or an efficient data annotation workflow, (3) a deep dive into your past projects or published research, where interviewers probe your understanding of your own work's limitations, trade-offs, and potential extensions, and (4) in some cases, a presentation of a relevant paper or research proposal. Interviewers are typically senior engineers and research scientists who work directly on Pony.ai's driving stack. Conversations are peer-to-peer and technically honest — they value candidates who can reason through ambiguity, acknowledge what they don't know, and articulate clear technical trade-offs. Culture fit at Pony.ai means demonstrating intellectual humility, comfort with fast iteration cycles, and genuine passion for the autonomous driving mission. Showing that you've studied Pony.ai's published research, blog posts, or technical talks signals authentic interest and preparation that interviewers notice.

What Pony.ai Looks For

  • Deep expertise in perception, computer vision, or deep learning — not surface-level familiarity, but the ability to discuss architectural choices, loss functions, and failure modes with precision
  • Strong C++ and Python proficiency with experience writing production-quality, performance-critical code for real-time systems
  • Published research in relevant ML or robotics venues (CVPR, NeurIPS, ICRA, ICLR, CoRL) — particularly valued for research intern and senior MLE roles
  • End-to-end ML engineering ownership: experience spanning data pipelines, model training, optimization (quantization, pruning, TensorRT), and deployment on edge hardware
  • Comfort with ambiguity and fast iteration — autonomous driving is a problem where requirements shift as technology capabilities evolve, and Pony.ai's teams move accordingly
  • Genuine passion for autonomous driving and its societal impact — interviewers at mission-driven companies like Pony.ai are attuned to candidates who care about the problem beyond the paycheck
  • Collaborative communication skills and the ability to explain complex technical work clearly — critical in a cross-functional, cross-cultural team spanning US and China offices

Frequently Asked Questions

How long does it typically take to hear back after applying to Pony.ai?
Response timelines at growth-stage AV companies like Pony.ai can vary, but many applicants report hearing back within 1-3 weeks for roles where their background closely matches the job description. For highly specialized positions like ML Runtime & Optimization or Perception Infrastructure, the recruiter may take additional time to consult with the hiring manager before reaching out. If you haven't heard back after three weeks, a brief follow-up email to the recruiter or a LinkedIn message expressing continued interest is appropriate and generally well-received. Keep in mind that Pony.ai operates across multiple time zones (US and China), which can affect communication cadence.
Does Pony.ai require a cover letter with my application?
Workable's application forms for Pony.ai may include an optional cover letter field, and while it's not always explicitly required, submitting one can differentiate you — especially for roles where your resume doesn't perfectly map to the job description. Use the cover letter to briefly explain why you're drawn to autonomous driving as a domain, how your specific technical background applies to the role, and any context that your resume can't convey (e.g., a career pivot from robotics hardware to ML software). Keep it under 300 words and make every sentence specific to Pony.ai — generic cover letters add no value and can signal low effort.
What degree level does Pony.ai expect for engineering and research roles?
Pony.ai's active postings commonly specify Master's or PhD requirements, particularly for research internships and ML engineering roles. This is typical in the autonomous driving industry, where the technical depth required — perception algorithms, deep learning architectures, real-time optimization — aligns with graduate-level training. However, candidates with a Bachelor's degree and substantial industry experience in relevant domains (computer vision, ML infrastructure, robotics) should not self-select out, as equivalent practical experience is often considered. Review each posting's specific education requirements carefully, as they vary by role and seniority level.
What programming languages and frameworks should I emphasize for Pony.ai roles?
C++ and Python are the two core languages across nearly all of Pony.ai's engineering roles, reflecting the AV industry's need for both high-performance runtime code and rapid ML experimentation. Beyond languages, emphasize specific frameworks and tools mentioned in their job descriptions: PyTorch for model development, TensorRT and ONNX for inference optimization, and CUDA for GPU programming. If you have experience with ROS (Robot Operating System), Eigen, or OpenCV, include those as well — they're commonly used in autonomous driving perception stacks. The key is specificity: 'PyTorch (custom C++ extensions, distributed training across 8 GPUs)' is far more compelling than 'familiar with PyTorch.'
How should I prepare for Pony.ai's technical interviews?
Preparation should span three areas. First, sharpen your algorithmic coding skills in C++ and Python — expect problems at LeetCode medium-to-hard difficulty, potentially with a geometric or spatial reasoning twist relevant to autonomous driving. Second, study ML system design: practice articulating how you'd build a perception pipeline end-to-end, including data ingestion, model training, optimization for edge deployment, and monitoring in production. Third, prepare a deep walk-through of your most impactful past project or publication — Pony.ai interviewers commonly probe the limits of your own work, asking about failure cases, trade-offs you made, and what you'd do differently. Reviewing Pony.ai's publicly available research and technical blog posts also helps you speak their language.
Does Pony.ai offer remote work options?
Autonomous driving development typically requires significant on-site presence due to the nature of the work — physical vehicles, sensor hardware, and on-vehicle testing are core parts of the development cycle. Based on available information, Pony.ai's US engineering roles are primarily based at their Fremont, California office. Some ML research and infrastructure work may offer more flexibility, but candidates should generally expect an on-site or hybrid arrangement rather than fully remote work. Check individual job postings for specific location requirements, and ask the recruiter during your initial conversation for the most current policy.
Can I apply to multiple roles at Pony.ai simultaneously?
Yes, Workable supports applying to multiple positions within the same company, and each application is tracked independently. If your background genuinely spans multiple roles — for example, if you're qualified for both the ML Runtime & Optimization position and the Software Engineer - Perception Infrastructure role — applying to both is reasonable. However, tailor your resume for each submission to match that role's specific keywords and requirements. Submitting the same generic resume to every open role can actually work against you, as it signals a lack of focus to recruiters who review applications across positions.
What makes Pony.ai different from other autonomous driving companies to work for?
Pony.ai distinguishes itself through several factors: its founding team's pedigree from Baidu's autonomous driving division, its dual US-China operational footprint that provides exposure to two of the world's largest AV markets, and its post-IPO status on NASDAQ which offers employees the financial upside of equity in a publicly traded company with the agility of a startup-sized engineering team. The company's partnership with Toyota adds manufacturing credibility, and its pursuit of Level 4 full autonomy (rather than incremental driver-assistance features) attracts engineers motivated by the hardest unsolved problems in the field. For candidates choosing between AV companies, the combination of research depth, real-world deployment scale, and lean team sizes — where individual engineers carry meaningful ownership — is a compelling differentiator.
How does Workable handle my application data at Pony.ai?
When you submit your application through Workable, the system parses your resume to extract structured data — your name, contact information, work history, education, and skills — and stores it in a searchable candidate profile. Pony.ai's recruiters can then filter and search candidates by keyword, degree, location, and other criteria. To ensure your profile is complete and accurate, review the parsed output if Workable gives you a preview, and correct any extraction errors before final submission. Your data is retained according to Pony.ai's privacy policy, so you can typically request deletion if needed. Submitting a cleanly formatted PDF minimizes parsing errors and ensures your profile accurately represents your qualifications.

Sample Open Positions

Check Your Resume Before Applying → View 4 open positions at Pony.ai

Related Resources

Similar Companies

Related Articles

Sources

  1. Pony.ai Careers - Open Positions — Workable / Pony.ai
  2. Pony.ai Company Overview and Technology — Pony.ai
  3. Pony.ai Glassdoor Reviews and Interview Insights — Glassdoor
  4. Workable ATS - How It Works for Applicants — Workable