AI Engineer Skills Companies Want in 2026: 3,449-Posting Analysis

The AI Engineer Title Has Settled Around the LLM Stack

Two years ago, "AI Engineer" was a fuzzy keyword that could mean almost anything: an ML researcher, a data scientist with a Python script, a backend engineer who fine-tuned a model once. In 2026 it has settled into a much more specific job: take a foundation model, wrap it in retrieval, monitoring, and an API, and ship it into a product. The variance lives in which model provider, which vector store, and which orchestration framework, not in what the work is.

To put numbers on it, we looked at every active AI Engineer posting on the InterviewStack.io job board as of May 2026, 3,449 listings, with skills extracted from descriptions and synonyms collapsed (so gen ai and generative ai count once, gcp and google cloud count once).

The headline: an AI Engineer posting in 2026 is, on average, a Python job plus an LLM job plus a retrieval job plus a cloud job rolled into one. Two skills appear in roughly two-thirds of postings or more, the RAG-plus-LangChain pattern has crossed the common-tier line, and a quiet salary premium has attached itself to anyone who can also handle the distributed-systems work behind those applications.

Key findings

• 3,449 active AI Engineer postings analyzed across the live job board as of May 2026.
• Python (71%) and LLMs (66%) are the only two table-stakes skills; 1,821 postings (53%) ask for both together.
• The LLM application stack has moved from differentiator to common: RAG (40%), Generative AI (39%), LangChain (25%), and OpenAI (20%) all now sit in the 20-50% common tier.
• Median US base salary is $146,000 (n=636), one of the highest role medians on our board.
• Distributed-systems and data-platform skills carry the biggest salary premiums: Distributed Systems ($180K, +$34K), Kafka ($171,500, +$25.5K), Apache Spark ($170K, +$24K), and Snowflake ($170K, +$24K).
• Only 6% of postings are entry-level (206 of 3,449); senior plus staff roles together make up 40% of the market.
• The US is 36% of postings, India is 13%: a much US-heavier mix than the Data Engineer market, where India is 23%.
• Onsite is still the default at 50% of postings; 34% are hybrid and 27% are remote (postings can carry multiple tags).

What Skill Families Define an AI Engineer Role in 2026?

Group every individual skill into the higher-level family it belongs to and count how many postings ask for at least one skill in that family. The role's actual shape emerges as a stack, not a single specialty, with the LLM application layer sitting on top of a software-engineering and cloud foundation.

Share of AI Engineer postings that ask for at least one skill in each family. A posting that mentions both PyTorch and TensorFlow counts once under "Machine Learning & AI".

The families that actually define the role:

1. Machine Learning & AI: 87% (LLMs, generative AI, machine learning, PyTorch, MLOps, TensorFlow, NLP, deep learning, computer vision, scikit-learn, pandas)
2. LLM Application Stack: 86% (RAG, LangChain, OpenAI, APIs, observability, vector databases, embeddings, Bedrock, FastAPI, scalability, microservices, distributed systems)
3. Coding Languages: 75% (overwhelmingly Python, with TypeScript, Java, and JavaScript as secondary languages)
4. Tools & Infrastructure: 67% (automation, monitoring, Docker, Kubernetes, Git, GitHub)
5. Cloud Platforms: 47% (AWS, Azure, Google Cloud)
6. Statistics & Experimentation: 37% (A/B testing, statistics)
7. Data Engineering Foundations: 29% (data pipelines, Apache Spark)
8. Querying & SQL: 24% (almost entirely SQL itself)

The "LLM Application Stack" family is what makes this role distinct. It bundles the skills you only see clustered together on postings that are productionizing LLMs: retrieval-augmented generation, vector stores, embeddings, the model-provider SDKs (OpenAI, Bedrock), and the operational skills (observability, scalability, distributed systems) needed to run them at scale. Two years ago this cluster did not exist as a coherent stack; today it appears in roughly the same share of postings as the underlying ML family.

The smallest families are also informative. Modern Data Stack sits at 15% and Data Visualization & BI at 15%, the inverse of what a Data Engineer posting looks like. The AI Engineer is rarely expected to build dashboards or own the warehouse; they consume the data once it lands. Read alongside the Data Engineer skills analysis, the contrast is sharp: the engineer stack centers on pipelines, warehouses, and orchestration; the AI Engineer stack centers on models, retrieval, and inference APIs.

What Are the Three Tiers of Individual AI Engineer Skills?

Drill into individual skills inside those families and three tiers emerge.

Top individual skills in AI Engineer postings, by share of listings that mention them. Skills above 50% are table stakes; 20-50% are common; 5-20% are differentiators. The data normalizer splits "llm" and "llms" into two buckets, which is why the chart shows both; in practice they refer to the same concept.

Table Stakes (50%+ of postings)

These appear in more than half of all AI Engineer postings. If your resume cannot credibly demonstrate them, you are filtered out before a recruiter reads a line.

• Python: 71% (AI Engineer + Python openings)
• LLMs: 66% (AI Engineer + LLM openings)

The table-stakes set is unusually narrow: just Python and LLMs. There is essentially no AI Engineer job in 2026 that does not involve writing Python that calls or fine-tunes a large language model. The two appear together in 1,821 postings, or 53% of the entire market, the single most common skill pair in the dataset and the closest thing to a canonical AI Engineer stack.

Worth noting: classical ML tooling like scikit-learn (8%) and even PyTorch (23%, common-tier) sit well below the table-stakes line. The role is no longer a research-ML role; it is an application role where the model is mostly a given and the engineering is around it.

Common Expectations (20-50% of postings)

This is where the LLM application stack lives.

• RAG (Retrieval-Augmented Generation, fetching context from a vector store and feeding it to an LLM): 40% (AI Engineer + RAG openings)
• Generative AI: 39%
• Machine Learning: 38%
• AWS: 37% (AI Engineer + AWS openings)
• Automation: 34%
• APIs: 33%
• Azure: 33%
• Monitoring: 31%
• Google Cloud: 25%
• LangChain (an open-source framework for chaining LLM calls, prompts, and tools): 25% (AI Engineer + LangChain openings)
• CI/CD: 24%
• A/B Testing: 24%
• PyTorch: 23%
• OpenAI: 20%

This tier is dominated by the LLM application stack. RAG (40%), LangChain (25%), and the OpenAI SDK (20%) all sit firmly in common-tier territory, a transition that happened over the last 18 months. As recently as 2024, RAG and LangChain were resume differentiators; they are now baseline expectations on a serious AI Engineer posting.

The cloud picture is similar to the rest of tech: AWS leads at 37%, Azure at 33%, Google Cloud at 25%. A candidate fluent in any one of the three is in the running for most postings; a candidate fluent in zero of them is struggling on roughly half the market, since most AI Engineer roles ship to production on a managed-cloud LLM service or hosted inference endpoint.

The A/B Testing entry (24%) is the most surprising line in the common tier. AI Engineers are increasingly responsible for measuring the impact of the systems they build, not just shipping them, so postings now bundle experimentation into the job description.

Differentiators (5-20% of postings)

These show up in a minority of postings but signal a more specialized, and, as we will see, often better-paid role.

• SQL: 19%
• Observability: 19%
• Vector Databases: 18%
• MLOps: 18%
• Data Pipelines: 18%
• Docker: 18%
• TensorFlow: 17%
• Kubernetes: 16%
• NLP: 15%
• Scalability: 15%
• Data Visualization: 13%
• TypeScript: 13%
• Embeddings: 12%
• Deep Learning: 12%
• Agile: 12%
• Statistics: 12%
• Java: 10%
• Git: 9%
• Prototyping: 9%
• Containerization: 8%
• JavaScript: 8%
• scikit-learn: 8%
• Microservices: 8%
• FastAPI (a Python framework for building production APIs): 8%
• Databricks: 7%
• React: 7%
• Computer Vision: 7%
• Distributed Systems: 7%
• Bedrock (AWS's hosted-LLM service): 7%
• GitHub: 7%
• Apache Spark: 6%
• pandas: 6%
• System Design: 6%

The infrastructure tier (Kubernetes, Docker, distributed systems, observability, MLOps) sits between 7% and 19%. None of them are required for most AI Engineer roles, but they are the skills that separate "AI Engineer who can run a demo" from "AI Engineer who can ship a production LLM application that stays up and gets debugged when it does not."

The vector-databases line (18%) is the cleanest signal in the data that the role has gotten serious about retrieval. Two years ago, almost no posting named a vector store; today, nearly one in five does, and the cluster of vector databases plus embeddings plus RAG appears almost exclusively together. If you are early in your career, learning one vector store deeply (most teams pick pgvector, Pinecone, Weaviate, or Qdrant) is a high-leverage move.

Which AI Engineer Skills Pay More Than the Baseline?

Salary numbers below are restricted to US postings only (where wage-transparency laws produce consistent disclosure) so they are directly comparable. The numbers are base salary: equity, bonuses, RSUs, and sign-on are not disclosed in postings, so total compensation at top employers is meaningfully higher than what we report here, especially at AI-native labs and frontier-model companies.

The overall median US base salary for AI Engineer postings is $146,000 (n=636). That is roughly $17,700 above the comparable median for Data Engineer postings ($128,300) and about $58,800 above the Data Analyst median ($87,200), a real, structural premium for the role's higher applied-AI and systems bar.

Median US base salary in USD for postings that mention each skill, among US AI Engineer postings with structured salary data.

The top-paying skills cluster around distributed systems and the data-platform layer, not the LLM application skills themselves. Skills with the largest premiums above the $146,000 baseline:

• Distributed Systems: $180,000 (n=41), about $34,000 above baseline
• Kafka: $171,500 (n=27), about $25,500 above baseline
• Apache Spark: $170,000 (n=53), about $24,000 above baseline
• Snowflake: $170,000 (n=51), about $24,000 above baseline

A second cluster of platform and applied-AI skills sits at roughly $150,000, premiums in the $4K to $7K range above baseline:

• React: $152,800 (n=41), about $6,800 above baseline
• A/B Testing: $152,400 (n=166), about $6,400 above baseline
• Observability: $151,800 (n=117), about $5,800 above baseline
• MLOps: $150,300 (n=88), about $4,300 above baseline
• LLMs: $150,000 (n=421), about $4,000 above baseline
• Computer Vision: $150,000 (n=41), about $4,000 above baseline
• Monitoring: $150,000 (n=192), about $4,000 above baseline
• Embeddings: $150,000 (n=87), about $4,000 above baseline
• Scalability: $150,000 (n=82), about $4,000 above baseline
• Microservices: $150,000 (n=41), about $4,000 above baseline
• System Design: $150,000 (n=31), about $4,000 above baseline

Skills that sit right at or slightly below baseline include LangChain ($145,000, n=139), PyTorch ($145,000, n=139), TensorFlow ($145,000, n=100), AWS ($145,000, n=245), Google Cloud ($145,000, n=150), Databricks ($145,000, n=56), and RAG ($147,100, n=245). The foundation skills (Python at $143,000, n=415; Generative AI at $140,000, n=250; Machine Learning at $140,000, n=249) actually sit a few thousand below baseline, the classic "every posting asks for this, so it does not differentiate" pattern.

The pattern is clear. The LLM application skills (RAG, LangChain, OpenAI) are now so common that they no longer carry a salary premium; they are required, not differentiating. The premium has shifted to the layer beneath: the engineers who can also handle distributed systems, streaming (Kafka), large-scale compute (Spark), and the data platform (Snowflake) are the ones who get paid for the combination. That is the practical signal in the data: an AI Engineer who can also do the heavy data-platform work earns roughly $20K to $30K more than one who only does the application layer.

Our interview-prep courses cover the foundations across system design, distributed systems, and ML; the question bank is where you drill the topics that come up in onsite rounds for the higher-premium specialties.

What Is the Dominant AI Engineer Skill Stack?

We computed every two-skill co-occurrence among the top 25 skills to find the combinations that show up together more often than chance.

The strongest pairs by lift, where lift greater than 1 means the two skills appear together more often than their individual frequencies would predict:

Skill pair	Postings that mention both	% of postings	Lift
LLMs + RAG	1,228	36%	1.34
LangChain + LLMs	768	22%	1.33
Generative AI + RAG	718	21%	1.33
AWS + RAG	711	21%	1.38
Python + PyTorch	731	21%	1.28
LangChain + Python	781	23%	1.26
APIs + LLMs	921	27%	1.23
CI/CD + Python	710	21%	1.21
Google Cloud + Python	744	22%	1.20
AWS + Python	1,082	31%	1.18
Python + LLMs	1,821	53%	1.12

Each pair tells you something concrete about how postings actually compose skills:

• LLMs + RAG (lift 1.34) is the strongest "what is the work" pair in the dataset. Postings that mention LLMs are 34% more likely to also mention RAG than baseline, because the dominant AI Engineer pattern in 2026 is not "fine-tune your own model" but "retrieve the right context and feed it to a foundation model".
• LangChain + LLMs (lift 1.33) and LangChain + Python (lift 1.26) signal how the work is built. Teams that adopt LangChain are looking for engineers who can chain LLM calls, tools, and retrieval steps in Python, not just write prompts.
• AWS + RAG (lift 1.38) is the highest-lift cloud pair. Companies on AWS are disproportionately the ones running production RAG systems, almost certainly because of Bedrock plus OpenSearch plus S3 plus Lambda forming a managed stack for it.
• Python + PyTorch (lift 1.28) marks the postings that still ask for model-level work alongside the application layer: fine-tuning, custom embedding models, or production inference in PyTorch rather than via a hosted API.
• APIs + LLMs (lift 1.23) and CI/CD + Python (lift 1.21) describe the productionization layer: postings that mention LLMs are 23% more likely to also ask for API design, and pipeline-as-code discipline is bundled with the role at almost the same rate.
• Python + LLMs (lift 1.12) is the dominant base stack. With 1,821 postings asking for both, Python + LLM AI Engineer roles make up 53% of the entire market, the closest thing to a single canonical AI Engineer stack.

The pattern: companies want a base layer (Python plus LLMs), a retrieval layer (RAG plus a vector store), an orchestration layer (LangChain or equivalent), a productionization layer (APIs, CI/CD, monitoring), and a cloud (AWS, Azure, or GCP). The "prompt engineer" role that some 2023 postings tried to describe does not exist in AI Engineer hiring; the role is a full-stack production engineer who happens to specialize in LLM applications.

Who's Hiring at Which Seniority Level?

We tagged each posting's seniority based on title keywords (Senior, Lead, Principal, Junior, Intern). Postings with no explicit signal default to mid-level.

Seniority distribution of AI Engineer postings.

• Mid-level: 54% (1,874 postings)
• Senior: 22% (747) (senior AI Engineer openings)
• Staff / Lead / Principal: 18% (622)
• Entry: 6% (206)

Two things stand out. First, only 6% of postings are explicitly entry-level, a narrower door than Data Analyst hiring (8% entry-level) but a wider one than Data Engineer hiring (3% entry-level). The bar for AI Engineer entry is higher than it sounds in the AI hype cycle, because most companies expect candidates to have shipped at least one applied-LLM or ML project somewhere first. Backend engineers, ML engineers, and data engineers transitioning in have an easier time than career-switchers from non-coding roles.

Second, the senior-and-above tiers (senior plus staff) are 40% of all postings. There is real career runway on the IC track, with substantial demand for staff-level engineers who can architect retrieval systems and inference platforms rather than just glue LLMs together. If you are targeting senior or staff AI Engineer roles, expect the differentiator skills (distributed systems, MLOps, observability, Kafka, Spark) to be required, not optional.

Where Are AI Engineer Jobs Located, and How Remote-Friendly Are They?

Geography for AI Engineer roles is much more US-concentrated than Data Engineer hiring, where India makes up nearly a quarter of postings. AI Engineer hiring is still flowing primarily through US tech and consulting.

Top countries by share of AI Engineer postings.

• United States: 36% (US-only AI Engineer openings)
• India: 13%
• United Kingdom: 5%
• Canada: 5%
• Germany: 4%
• Singapore: 3%
• France: 2%
• Spain: 2%
• Poland: 2%
• Netherlands: 2%

The US is the dominant single market at more than a third of all postings. India is a distant second at 13%, roughly half its share of the Data Engineer market, a reflection of how much AI Engineer demand still concentrates in US-based product companies, AI labs, and the major consultancies' US delivery practices. The Singapore line (3%) is notable; it punches above its weight relative to most other tech roles, driven by a cluster of AI hiring at regional tech firms and a research university (Nanyang Technological University) that shows up in the top employers below.

The "AI Engineer is a perfect remote-first role" assumption is partly true, but onsite still leads.

Share of AI Engineer postings tagged with each work mode. Some postings carry multiple tags (e.g., "Hybrid or Remote"), so percentages sum to more than 100%.

• Onsite: 50% of postings (1,735)
• Hybrid: 34% (1,157)
• Remote: 27% (926) (fully-remote AI Engineer openings)

Postings can carry multiple work-mode tags when a company says "Hybrid or Remote", which is why the percentages sum to more than 100%. Fully remote AI Engineer roles do exist and are comparable in share to Data Engineer roles (27% in both), but the dominant mode is still onsite. The remote share concentrates in AI-native startups and product-led tech companies; financial services, consulting, and government default to onsite or hybrid.

Who's Hiring AI Engineers in 2026?

The top hiring companies on our board mix Big Four consulting, AI-native product companies, frontier-model labs, enterprise software, and a meaningful staffing-and-aggregator tail.

Top companies by active AI Engineer postings. Counts include all locations of the same job.

• PricewaterhouseCoopers: 97 postings (Big Four consulting)
• Hyphen Connect Limited: 45 (staffing and recruiting)
• EverAI: 38 (AI product company)
• Jobgether: 35 (job aggregator and staffing)
• NVIDIA: 32 (AI hardware and platforms)
• Nanyang Technological University: 27 (academic research)
• Celonis SE: 27 (process-mining enterprise software)
• Sezzle: 26 (fintech)
• Huawei Technologies Canada: 24 (telecom research)
• Exadel: 21 (software services)
• Accenture: 21 (global consulting)
• Micron Technology: 20 (semiconductors)

A few names worth flagging further down the top 20: Mistral AI (16 postings) is the only frontier-model lab in the list, AstraZeneca (16) and Royal Bank of Canada (15) represent the pharma and banking pull into AI Engineering, and Nebius Academy (17) shows the training and education segment building out its own AI Engineering teams.

A few of the highest-volume entries (Hyphen Connect, Jobgether, Jack & Jill, Nexthire) are staffing and aggregator brands that re-post roles for many client companies, which is why their counts run high; the direct-employer leaders on the list are PwC, NVIDIA, EverAI, Celonis, Sezzle, Huawei Canada, Accenture, Micron, Mistral AI, AstraZeneca, and RBC.

The shape of the list confirms two things the rest of the data already suggested. First, a meaningful share of AI Engineer demand still flows through consulting firms, not direct posts from end employers; PwC, Accenture, Capco, and Booz Allen Hamilton together account for more than 150 listings. Second, the rest is genuinely diverse: AI-native product companies (EverAI, Mistral AI), AI infrastructure (NVIDIA, Micron), enterprise software (Celonis), fintech (Sezzle, RBC), pharma (AstraZeneca), and academia (Nanyang). There is no single industry that dominates AI Engineer hiring, which is unusual for a fast-growing role. For specific company processes, our interview preparation guides break down the rounds, topic priorities, and behavioral expectations company by company.

How to Use This in Your Job Search

If you are preparing for an AI Engineer job hunt, the data points to a clear sequence.

1. Build the two table-stakes skills ruthlessly. Python and LLM application work are the two filters every posting applies. Not weekend-tutorial Python, production Python: writing testable modules, handling errors, packaging code that runs reliably behind an API. Not "I ran a ChatGPT prompt", LLM application work: building a real retrieval system, evaluating outputs, handling latency, debugging hallucinations in production. The two appear together in 53% of postings, the single most common pair in the dataset.

2. Pick a cloud, a vector store, and a framework. AWS is the largest single cloud at 37%, with the strongest tie to RAG (lift 1.38), but Azure (33%) and Google Cloud (25%) cover comparable ground in their respective company segments. For retrieval, pick one vector store and learn it deeply (pgvector, Pinecone, Weaviate, or Qdrant are the common picks); the postings cluster vector databases (18%) and embeddings (12%) together, so the two skills are bundled. For orchestration, LangChain is the safest default at 25% of postings, but understanding the underlying pattern (chains, tools, retrievers) matters more than the specific library.

3. Add a differentiator from the platform layer. The salary data is unambiguous: the skills companies pay the largest premiums for are not the LLM application skills themselves but the platform layer beneath them. Distributed Systems, Kafka, Apache Spark, and Snowflake each move your median US base salary by $24K to $34K over the role baseline. Pick one that fits the kind of system you want to build (high-throughput streaming, large-scale compute, or warehouse-native data) and learn it deeply enough to talk through trade-offs in an onsite.

4. Drill the topics, then practice the rounds. Reading about AI Engineer skills is easy; performing under interview conditions is the hard part. Our interactive courses cover the foundations across system design, statistics, and applied ML. The question bank lets you drill ML, system design, distributed systems, and LLM application topics one at a time. AI mock interviews let you practice the full round under realistic conditions, with on-demand feedback on system-design and ML-design questions specifically.

5. Filter the job board for your stack. Browse current AI Engineer openings on the InterviewStack.io job board and combine role and skill filters to narrow to your exact stack, e.g., AI Engineer + RAG + AWS or AI Engineer + LangChain + Python. The board updates daily, so the listings are current.

FAQ

Q. What skills do companies want for AI Engineer roles in 2026?

Python and LLMs are table stakes, appearing in 71% and 66% of postings respectively. Above that base, RAG (40%), Generative AI (39%), Machine Learning (38%), AWS (37%), automation (34%), APIs (33%), Azure (33%), monitoring (31%), Google Cloud (25%), LangChain (25%), CI/CD (24%), A/B Testing (24%), PyTorch (23%), and OpenAI (20%) sit in the common tier. Vector databases, MLOps, observability, and distributed systems are differentiator skills that pay real premiums.

Q. What is the median salary for an AI Engineer in 2026?

The median US base salary across 636 AI Engineer postings with disclosed salary data is $146,000. That figure excludes equity, bonuses, and sign-on, so total compensation at top employers runs meaningfully higher, especially at AI-native labs and frontier-model companies.

Q. Which AI Engineer skills pay the highest premium over the role baseline?

Among US postings, the largest premiums attach to distributed-systems and data-platform specialties. Distributed Systems ($180,000, +$34K over the $146,000 baseline), Kafka ($171,500, +$25.5K), Apache Spark ($170,000, +$24K), and Snowflake ($170,000, +$24K) lead the table. A cluster of platform and applied-AI skills follows at roughly $150,000 (+$4K to +$7K): React, A/B Testing, Observability, MLOps, LLMs, Computer Vision, Monitoring, Embeddings, Scalability, Microservices, and System Design.

Q. Is AI Engineer a good entry-level role to break into?

Entry-level access is narrow but not closed. Only 6% of AI Engineer postings are explicitly entry-level (206 of 3,449), compared with 8% for Data Analyst and 3% for Data Engineer. Most companies expect candidates to have already shipped at least one applied-LLM or ML project, so career switchers typically route through ML engineering, backend, or data-engineering roles before stepping in.

Q. Where are most AI Engineer jobs located, and how remote-friendly are they?

The United States is the largest single market at 36% of postings, followed by India at 13%, the UK (5%), Canada (5%), Germany (4%), and Singapore (3%). About 27% of postings are tagged remote, 34% hybrid, and 50% onsite (some postings carry multiple tags), so onsite remains the dominant default.

Q. Which companies hire the most AI Engineers in 2026?

The top of the list mixes Big Four consulting, AI-native companies, and enterprise software: PricewaterhouseCoopers (97 active postings), Hyphen Connect Limited (45), EverAI (38), Jobgether (35), NVIDIA (32), Celonis SE (27), Nanyang Technological University (27), Sezzle (26), Huawei Technologies Canada (24), Exadel (21), Accenture (21), and Micron Technology (20). Mistral AI, AstraZeneca, and Royal Bank of Canada also appear in the top 20.

Q. What is the dominant AI Engineer skill stack in 2026?

Python plus LLMs is the foundation, appearing together in 1,821 postings (53% of the market) with a co-occurrence lift of 1.12. The most over-represented combinations layer RAG, LangChain, and PyTorch on top of that base: LangChain + LLMs (lift 1.33), Python + PyTorch (1.28), LangChain + Python (1.26), APIs + LLMs (1.23), and CI/CD + Python (1.21) all describe stacks built around productionizing LLM applications.

Final Thoughts

The AI Engineer role in 2026 has settled into a coherent stack (Python plus LLMs plus retrieval plus a cloud) with a real ladder above it and one of the highest role medians on our board. The trade-off is that the application-layer skills (RAG, LangChain, OpenAI) are now so common that they no longer differentiate; the salary premium has migrated to the platform layer underneath. If you can route through a backend, ML, or data-engineering role to build the production and distributed-systems reps, the senior and staff tier opens up quickly, and the differentiator skills compound from there.

We will refresh this analysis quarterly so the trend lines stay current.