Reka AI Internship 2027–2028: Programs, Deadlines & How to Apply
Last updated: July 2026
Reka AI reached unicorn status with a $1B valuation after its $110M Series B led by NVIDIA and Snowflake in July 2025, all while keeping a team of roughly 60 people spanning 17 nationalities. Founded by former DeepMind and Meta AI researchers, the company builds natively multimodal foundation models that process video, image, audio, and text within a single framework. There is no formal internship program, which means every opening is filled through direct outreach, academic networks, or ad hoc postings on their Ashby job board. For PhD students and exceptional Master's candidates with multimodal ML research backgrounds, that informality is the opportunity.
Quick Facts
| Fact | Detail |
|---|---|
| Where to apply | reka.ai/careers and jobs.ashbyhq.com/reka. No formal intern postings exist as of July 2026; cold outreach to research leads is the most effective path |
| Application window (2027–28) | No official timeline. Estimated November 2026 through February 2027 for summer 2027 roles, based on AI startup hiring patterns. Rolling review expected |
| Rolling? | Likely yes. Reka does not run a formalized intern program; any openings are sourced through direct outreach, academic connections, or ad hoc postings, which implies rolling review |
| Eligibility | PhD student (strongly preferred) or exceptional Master's in CS, ML, or related field. Published papers in top ML/AI venues (NeurIPS, ICML, ICLR, CVPR, ACL) expected |
| Duration | Estimated 10 to 12 weeks for summer internships based on industry norms. Research internships at similar labs can extend to 3 to 6 months |
| Compensation | No verified intern data. Estimated $8,000 to $12,000 per month based on comparable Bay Area AI startups at the unicorn stage. Full-time software engineers earn $200K to $280K+ total compensation |
| Visa sponsorship | Reka has sponsored H1B visas for full-time roles ($150K to $170K salary filings). CPT/OPT for US-based international students is likely. 17 nationalities represented across the team |
| Locations | Sunnyvale, CA (HQ), London/UK (research hub), Singapore. Most current job listings specify Remote |
| # Programs | No formal intern tracks. Estimated 3 to 5 possible areas based on current full-time roles: ML Research (Multimodal), Applied AI/ML Engineering, Robotics Research, Data Intelligence, and Inference/Systems |
Reka AI has no structured internship program. With roughly 60 employees and an estimated 0 to 5 intern slots per cycle, any position is filled ad hoc through cold outreach, academic referrals, or occasional Ashby postings. PhD students with published multimodal ML research are the primary target.
Externships are short, remote projects where you finish real work for a real company. The Wayfair AI Agent Engineering Externship and Beats by Dre Data Analytics Externship build the Python, ML engineering, and data analysis evidence that a cold outreach to Reka's research team needs to stand out. Explore all Externships.
What Is a Reka AI Internship?
A Reka AI internship is an ad hoc, research-driven placement at a frontier multimodal AI lab with roughly 60 employees across three continents. There is no structured intern program, no cohort, and no public application timeline. Instead, Reka fills a handful of intern slots each cycle through direct outreach, academic referrals, and occasional postings on its Ashby job board. The company was founded in late 2022 by former researchers from Google DeepMind and Meta AI, including pioneers behind Gemini and Veo, and has raised $170M across two rounds to build models like Reka Flash, Core, and Edge. Interns work alongside a small team of world-class researchers on problems ranging from multimodal architecture design to inference optimization. The flat organizational structure, with every researcher holding the title Member of Technical Staff, means there is no junior tier to hide behind. You are expected to contribute publishable-quality research from the start.
When Do Reka AI Internship Applications Open for 2027–2028?
Reka AI does not publish an internship application timeline, and there is no structured hiring cycle to reverse-engineer. The company fills positions ad hoc, which means the timeline you follow is the one you create through proactive outreach. Based on AI startup hiring patterns and Reka's current team size of roughly 60 people, the best window for cold outreach is September through December 2026, three to six months before a summer 2027 start. Any postings that appear will likely show up on the Ashby job board between November 2026 and February 2027 and close on a rolling basis. The critical insight: at a lab this small, a well-timed cold email to a research lead with a relevant publication attached outperforms any job board application.
No Reka AI internship postings are live. This is the build window: strengthen your multimodal ML research portfolio, draft a cold-outreach email to Reka's research leads, and publish or submit papers to top venues before outreach season begins in fall 2026.
Prime cold-outreach window. Contact Reka research leads and founders 3 to 6 months before your desired start date. Include your research background, 1 to 2 relevant publications, and specific interest in Reka's multimodal work. Academic referrals through DeepMind or Google Brain alumni networks carry weight.
If Reka posts ad hoc internship roles, they will likely appear on the Ashby job board during this window. Monitor jobs.ashbyhq.com/reka and apply immediately when a relevant listing appears. Rolling review means early applicants have the best odds.
Interview phase. Expect a recruiter screen (30 minutes), a technical coding round focused on Python and PyTorch (45 to 60 minutes), a research discussion or project deep-dive (45 to 60 minutes), and possibly a conversation with a co-founder.
Estimated 10 to 12 weeks, June through August. Work alongside a flat team of roughly 60 researchers on multimodal foundation models, inference optimization, or robotics research. Strong performers at a lab this size are natural candidates for full-time Member of Technical Staff roles.
Why You Must Apply the Week Applications Open
There is no week one at Reka AI in the traditional sense, because there is no structured application window. The advantage goes to candidates who reach out before a role is ever posted. With roughly 60 employees and an estimated 0 to 5 intern slots per cycle, Reka fills positions through direct outreach and academic networks far more often than through job board applications. That means the most important move is a well-crafted cold email to a research lead, sent three to six months before your desired start date, with a relevant publication or project attached. A personal referral from someone in the DeepMind, Google Brain, or Meta AI alumni network multiplies your odds further. By the time an ad hoc posting appears on Ashby, the strongest candidates are often already in conversation.
Which Reka AI Internship Programs Should You Target?
Reka AI does not publicly list distinct internship programs. However, the company's current full-time Member of Technical Staff roles and product areas map to several clear research tracks. Which one you target should match your existing research expertise, because at a lab this small, there is no training ramp. You are expected to contribute from day one.
| Program | Focus | Duration | Key skills |
|---|---|---|---|
| ML Research Intern (Multimodal Models) | Training, architecture design, and modality alignment for Reka Flash, Core, and next-generation multimodal foundation models processing video, image, audio, and text | 10 to 12 weeks | Python, PyTorch, Transformer architectures, multimodal learning, CLIP-style encoders, distributed training |
| Applied AI / ML Engineering Intern | Building production ML pipelines and integrating models into Reka Vision and Reka Research platforms, with emphasis on inference optimization and deployment | 10 to 12 weeks | Python, PyTorch, model serving, API development, inference optimization, quantization |
| Robotics Research Intern | Working on Reka's robotics research track, including Claru training data infrastructure, egocentric video, and robotics trajectories | 10 to 12 weeks | Python, PyTorch, robotics fundamentals, computer vision, trajectory planning, world models |
| Data Intelligence Intern | Training data infrastructure, large-scale data curation, and quality systems for multimodal datasets | 10 to 12 weeks | Python, data engineering, distributed systems, data quality frameworks, large-scale data processing |
| Inference / Systems Intern | Working on the Infer engine with focus on KV-cache management, quantization, distillation, and latency optimization for model serving | 10 to 12 weeks | Python, C++, systems engineering, GPU/TPU optimization, model compression, low-latency serving |
Check all current openings on the Reka Ashby job board. As of July 2026, all listed positions are full-time Member of Technical Staff roles. No dedicated internship listings have been posted, which is typical for a lab of this size.
What Are the Eligibility Requirements?
Reka AI has not published formal internship eligibility criteria, but the pattern across its full-time hiring and the backgrounds of its team members makes the bar clear:
• Enrollment: currently pursuing a PhD (strongly preferred) or Master's in Computer Science, Machine Learning, or a related field. The founding team comes from Google DeepMind and Meta AI, and that pedigree sets the expectation.
• Research output: published papers in top ML/AI venues such as NeurIPS, ICML, ICLR, CVPR, ACL, or EMNLP, especially in multimodal learning, efficient training, or foundation models. First-author publications are a significant differentiator.
• Prior experience: a prior internship at a top AI lab (DeepMind, FAIR, Google Brain, OpenAI, Anthropic) is a strong signal but not a hard requirement. Demonstrated ability to train models at scale and contributions to open-source ML projects also carry weight.
• Work authorization: Reka has filed H1B visas at $150K to $170K salary for full-time employees, and 17 nationalities are represented across the team. CPT/OPT for US-based international students is likely supported. The UK entity and Singapore presence open additional paths for non-US candidates.
Does Reka AI Have a GPA Cutoff?
No GPA cutoff has been published or reported. Reka's hiring emphasis falls squarely on research output, technical depth, and the ability to work in a small, flat team. Published papers, demonstrated coding ability in Python and PyTorch, and a clear connection to Reka's multimodal research agenda matter far more than a transcript number.
What Skills Does Reka AI Look For, and How Do You Build Them?
Read across Reka's full-time Member of Technical Staff postings and interview reports, and the pattern is clear. Python and PyTorch appear in every technical role. Large-scale model training and Transformer architecture knowledge are assumed. Multimodal learning, the ability to work across vision, language, video, and audio within a unified framework, is what separates Reka from a generic ML lab. The interview process tests not just algorithmic fluency but ML system design: expect questions on inference platform architecture, KV-cache management, and modality alignment. On the soft side, Reka's flat team structure means self-direction and the ability to collaborate across research and engineering without a management layer are non-negotiable.
What Reka AI looks for in interns
Skills across 5 Reka AI intern & analyst job descriptions · 2025–26 Reka AI MTS job descriptions and interview data, projecting 2026–2027
Method: full-text analysis of five Reka AI Member of Technical Staff job descriptions and interview reports from TechInterview.org. Prior-cycle basis; Reka does not publish dedicated internship JDs, so skill requirements are inferred from full-time postings.
How Is Demand for ML Research Interns Moving Right Now?
Machine learning intern hiring right now: July 2026
Across US machine-learning-intern postings tracked this week · aggregate market data, all employers
The ML research niche within intern hiring pays more and grows faster than general SWE. At a frontier lab like Reka, you are competing for one of the very few slots that exist, but the skills that get you in translate directly to the broader market if you need a backup plan.
Method: aggregate analysis of US software-engineer-intern, machine-learning-intern, and software-engineer postings via Adzuna, July 2026 baseline. The sample indexes a fraction of all US postings, so figures show direction and relative level, not total market share.
Build These Skills Before You Apply
Every skill in that chart maps to a remote Externship where you finish a real project before cold outreach season.
| Skill (from real JDs) | JD evidence | Externship that builds it |
|---|---|---|
| Python, PyTorch & large-scale model training | MTS JDs: "deep proficiency in Python and PyTorch" and "experience with large-scale model training" | Wayfair AI Agent Engineering |
| Data analysis, experiment design & metrics | Applied AI JDs: "production ML pipelines, data curation, quality systems for large-scale multimodal datasets" | Beats by Dre Data Analytics |
| ML system design & inference optimization | Systems JDs: "KV-cache management, quantization, distillation, latency optimization" | Wayfair AI Agent Engineering |
The Wayfair project gives you agent-engineering work that produces the build-and-explain evidence a Reka cold outreach needs, and the Beats project delivers the data-analysis deliverable that demonstrates end-to-end thinking.
What Is the Reka AI Application and Interview Process Like?
Reka AI has no formalized internship application process. Based on its full-time interview structure and AI startup norms, the most likely path to an internship looks like this:
1. Identify target contacts. Research Reka's team on LinkedIn. Focus on research leads, hiring managers, and founders: Dani Yogatama (CEO), Yi Tay (Chief Scientist), Cyprien de Masson d'Autume (CTO), Qi Liu, and Mikel Artetxe. Monitor reka.ai/careers and jobs.ashbyhq.com/reka for any ad hoc intern postings.
2. Send a concise cold outreach. Email or message a relevant research lead. Include your research background, one to two relevant publications, specific interest in Reka's multimodal work, and availability for summer 2027. If your advisor or lab has connections to DeepMind, Google Brain, or Meta AI alumni, use those referral channels.
3. Recruiter or hiring manager screen (30 minutes). Initial conversation covering your background, motivation, and fit with Reka's research direction. Expect questions about why Reka specifically and what you would work on.
4. Technical coding round (45 to 60 minutes). Python and PyTorch focus. Medium-to-hard DSA problems, often ML-themed. You must state time and space complexity. Clean, PyTorch-compatible code is expected.
5. Research discussion and project deep-dive (45 to 60 minutes). Present your most relevant research. Expect questions on CLIP-style encoders, cross-attention mechanisms, modality alignment, and inference optimization. A research-track candidate may face an additional paper-discussion round.
So the two things to prepare are clear: Python and PyTorch fluency for the coding round (focus on ML-themed problems, not generic LeetCode), and a deep command of your own published research that you can connect directly to Reka's multimodal agenda. The team is small enough that a co-founder may join the final conversation, which means mission alignment and cultural fit carry real weight. And note: NTU Singapore has a partnership pipeline with Reka, so APAC-based students should explore that channel as well.
What Students on Reddit Say
Reka AI appears in several AI community discussions, though none focus specifically on internship experiences. Here are the most relevant threads, all paraphrased.
Reka is recognized as a legitimate frontier AI lab building natively multimodal models, with its Flash model praised for rivaling much larger models in quality while being optimized for fast, efficient workloads.
The AI community discusses how smaller labs like Reka compete against big tech by specializing in areas like multimodal understanding and efficient inference, carving out niches that the largest labs treat as secondary priorities.
Reka's founding team is noted among the wave of researchers who left DeepMind, Google Brain, and Meta AI to start independent labs, with the company listed alongside peers building efficient foundation models optimized for low-cost inference.
How Do You Stand Out at a Lab with 60 People and No Intern Program?
Three moves, all before a posting appears. First, cold-email a research lead with substance, not a generic "I am interested in ML" message: include one to two relevant publications, name the specific Reka product area you want to work on (Flash? Vision? Claru?), and explain what you would contribute in 10 to 12 weeks. At a lab this small, a targeted cold email outperforms any job board application. Second, publish in multimodal ML. Reka's founding team comes from DeepMind and Meta AI, and they evaluate candidates by research output, not transcripts. A first-author paper at NeurIPS, ICML, or ICLR on vision-language models or efficient training is the strongest signal you can send. Third, bring PyTorch fluency that goes beyond coursework. The technical interview tests ML-themed coding with PyTorch, and the research discussion probes architecture decisions on cross-attention, modality alignment, and inference optimization. Practice explaining your own code at the level of a peer review, not a tutorial.
What Other Companies Should You Consider?
Reka AI sits in the frontier multimodal AI lab tier. If you are building an AI research internship list, these are the direct neighbors:
- OpenAILargest frontier lab with structured research intern programs and broader product scopeGuide →
- AnthropicSafety-focused frontier lab with structured intern cohorts and constitutional AI researchGuide →
- Google DeepMindLargest pure research lab with extensive intern programs across ML, robotics, and multimodal AIGuide →
- Scale AIData infrastructure company for AI with ML engineering and research intern tracksGuide →
- NVIDIAGPU and AI compute leader with extensive research internships in deep learning and inference optimizationGuide →
Our tech internships summer 2027 guide maps the whole landscape, timeline by timeline.
FAQ
Does Reka AI have a formal internship program?
No. As of mid-2026, Reka AI does not advertise a structured internship program. With roughly 60 employees, any intern positions are filled ad hoc through direct outreach, academic connections, or occasional postings on the Ashby job board at jobs.ashbyhq.com/reka.
What kind of candidates does Reka AI look for?
Reka strongly prefers PhD students or exceptional Master's students with published research in multimodal ML, efficient training, or foundation models. First-author papers at top venues like NeurIPS, ICML, ICLR, or CVPR and prior experience at elite AI labs are significant differentiators.
How much does a Reka AI internship pay?
No verified intern compensation data is available. Based on comparable Bay Area AI startups at the unicorn stage, estimated monthly compensation is $8,000 to $12,000. Full-time software engineers at Reka earn $200K to $280K or more in total compensation.
Does Reka AI sponsor visas for interns?
Reka has sponsored H1B visas for full-time employees at $150K to $170K salary. For interns, CPT and OPT for US-based international students is likely. The company has 17 nationalities represented, indicating openness to global talent. UK and Singapore offices may offer alternative paths.
Where are Reka AI internships located?
Reka has offices in Sunnyvale, CA (HQ), London/UK (research hub), and Singapore. Most current job listings specify Remote as the location, so remote internships may be possible depending on the role.
How do I apply for an internship at Reka AI?
The most effective path is direct outreach. Cold-email research leads or founders with a concise pitch that includes your research background, relevant publications, and specific interest in Reka's multimodal work. Academic referrals through advisor networks connected to DeepMind or Google Brain alumni carry significant weight.
What programming languages and frameworks does Reka AI use?
Python and PyTorch are essential. Candidates should also have experience with large-scale model training, Transformer architectures, distributed training frameworks, and ideally multimodal learning across vision, language, video, and audio.
What is the Reka AI interview process like?
Based on full-time interview data, expect a recruiter screen (30 minutes), a technical coding round focused on Python and PyTorch with ML-themed problems (45 to 60 minutes), a research discussion or project deep-dive (45 to 60 minutes), and possibly a conversation with a co-founder. No standardized online assessment has been reported.
Reka AI has no formal intern program, which means every opening is won through direct outreach and research credentials. Spend the runway building proof: a remote Externship gives your cold email a finished, technical deliverable to point at when the outreach window opens in fall 2026.
About the Author
Bifei Wang has spent 17 years focused on human flow and the growth of young professionals, spanning international education, career training and coaching, and recruitment process outsourcing. Over 7 years at Extern, he has had one-on-one sessions with thousands of students exploring careers in consulting, finance, tech, marketing, and data, giving him a firsthand view of how the job market has shifted for early-career professionals and what it actually takes to break in.
