It’s Q2 2026, and a fast-growing fintech company in San Francisco is losing senior AI engineers. Their average time-to-hire is 67 days while competitors with clearer, faster processes extend offers in under 30. Three of their last five final-round candidates accepted competing offers before the hiring committee aligned on requirements. The root cause was a vague, recycled position description that meant different things to every interviewer.
This problem is common at tech companies scaling AI and engineering teams. Generic wish lists of technologies create misaligned expectations, inconsistent candidate evaluations, rejected offers, and onboarding friction because success in the first 90 days was never documented.
This article is for hiring managers, VPs of Engineering, Heads of Talent, and recruiters at fast-growing tech startups and scale-ups hiring AI, ML, and software engineers. A position description is a detailed, seat-specific document defining purpose, responsibilities, qualifications, and success metrics, different from a generic job description or an external job posting.
Key Takeaways
A position description is a detailed, role-specific document used to manage expectations, performance, and hiring quality, distinct from a broader job description template for a job family or level.
Fast-growing tech companies face urgent hiring challenges including long engineering hiring cycles, recruiters screening hundreds of resumes per role, and inconsistent candidate quality affecting most technical hires, making clear position descriptions essential.
An effective position description includes position summary, reporting structure, key responsibilities with measurable outcomes, required and preferred qualifications, tech stack, and compensation, and Fonzi’s multi-agent AI uses these structured documents to streamline screening, detect fraud, and support accurate candidate matching while keeping managers in full control.
Position Description vs. Job Description vs. Job Posting
These three documents serve different but connected purposes in a modern hiring stack. Understanding the distinction helps you create the right document for the right audience, and avoid the confusion that slows down technical hiring.
Dimension | Position Description | Job Description | Job Posting |
Audience | Internal (hiring manager, HR, interviewers) | Internal (HR, compensation teams) | External (job applicants, prospective employees) |
Level of Detail | High; specific systems, metrics, reporting lines | Medium; generic duties for a job family | Low to medium; marketing-focused highlights |
Ownership | Hiring manager + HR/Talent | Human resources / compensation | Recruiting / talent marketing |
Update Frequency | When role scope changes or annually | When job family is restructured | Per requisition or campaign |
Purpose | Define expectations, evaluate performance, guide interviews | Standardize leveling and compensation | Attract qualified candidates |
A job description is generic and role-based. It might describe what a “Software Engineer II” does across your organization, including common responsibilities, typical qualifications, and compensation bands. It helps HR ensure consistency across teams but lacks the specificity needed to hire for a particular position.
A position description is specific to a seat on a team. For example, “Backend Engineer – Payments Platform, SF, reporting to the Director of Platform Engineering.” It includes the exact systems this person will own, the problems they will solve in the next 12 months, and the success metrics their manager will use to evaluate performance. This level of detail helps interviewers assign work and evaluate candidates against consistent criteria.
A job posting is the externally-facing version distilled from your position description. It is tailored to attract candidates on channels like LinkedIn, Fonzi, or your careers page. Job postings are marketing documents. They should accurately reflect the role while selling prospective employees on why this opportunity matters.
In many fast-growing companies, position descriptions are underdeveloped or skipped entirely. Teams jump straight from a Slack message (“We need a senior ML person”) to a job posting. This creates confusion across recruiters, interviewers, and candidates, and it is one of the primary reasons technical hiring cycles stretch past 60 days.
Core Components of an Effective Position Description
A strong position description has repeatable components that serve both human recruiters and AI systems like Fonzi’s multi-agent models. Each element reduces ambiguity and helps everyone involved in hiring, from sourcers to final interviewers, develop a clear understanding of what the role requires.
Position Summary: A 3–4 sentence statement describing the role’s purpose and how it fits into the team and organization. This is not a list of tasks. It is the “why” behind the role. Start with what problems this person will solve and what impact they will have.
Reporting & Team Context: Who this position reports to, the size and structure of the team, and key cross-functional relationships. This helps candidates evaluate culture fit and interviewers calibrate seniority expectations.
Key Responsibilities: 5–7 bullets describing primary duties using action verbs in present tense. Each bullet should answer: “What will this person do, and what outcome will they drive?” Be specific such as “design and deploy ranking models to improve CTR by 15%” beats “work on ML projects.”
Outcomes & Success Metrics: What does success look like in the first 6–12 months? Quantified targets help candidates self-select and give managers a framework to evaluate candidates during interviews.
Required Qualifications: Non-negotiable minimum qualifications including education, years of experience, and specific duties the person must be able to perform on day one. Be honest; if 5 years of production ML experience is truly required, say so. If it’s flexible, move it to preference.
Preferred Qualifications: Nice-to-have skills and experience that would accelerate ramp time but aren’t deal-breakers. This is where you list adjacent technologies, domain knowledge, or leadership experience.
Tech Stack & Tools: For AI-heavy roles, this section is critical. List specific frameworks (PyTorch, TensorFlow), infrastructure (Kubernetes, AWS, GCP), data tools (Snowflake, Databricks), and any internal systems the person will use. Vague descriptions like “experience with cloud platforms” attract mismatched candidates.
Working Model & Time Zone Expectations: Remote, hybrid, or on-site? How many days in office? Which time zones must this person overlap with? These details prevent wasted cycles with candidates who can’t meet location requirements.
Compensation Range & Leveling: Increasingly required by law in states like California and New York, compensation transparency also improves candidate quality. Include the salary band, equity expectations, and how this role maps to your internal leveling framework.
Legal/Compliance Statements: Include EEO statements, physical requirements if applicable (especially for roles involving data center access), and any certifications or licenses required for the particular role.
For AI and engineering roles, especially LLM engineers, MLOps specialists, and Staff+ positions, the tech stack, data domains, and model responsibilities must be unusually specific. The difference between “experience with transformers” and “experience fine-tuning LLMs for production retrieval-augmented generation systems” is the difference between 200 unqualified applicants and 20 strong matches.
Step-by-Step: How to Write a Position Description (With AI Support)
Writing a position description doesn’t need to take 6 hours. Here’s a practical workflow for hiring managers and recruiters collaborating on a new engineering or AI role in 2026.
Step 1: Conduct a Quick Role Discovery
Before writing anything, answer these questions with the hiring manager:
What problems will this person solve in the next 12 months?
What will be different about our product/systems because this person is here?
Who will they work with most closely?
What does the work schedule look like, and where does this person need to be located?
This 30-minute conversation surfaces the real job duties required, not a recycled list from the last similar hire.
Step 2: Map Responsibilities to Measurable Outcomes
Transform vague statements into specific, measurable responsibilities. Follow this sentence structure:
[Action verb] + [what] + [for what purpose/outcome]
Examples:
“Design and implement feature extraction pipelines to reduce model training time by 40%”
“Own the end-to-end deployment process for recommendation models serving 10M+ daily active users”
“Collaborate with product to define success metrics for personalization experiments”
Use present tense throughout. Begin each bullet with an action verb that describes actual work and not ambiguous terms like “support,” “assist,” or “handle” unless paired with clear context.
Step 3: Define Non-Negotiable Skills vs. Nice-to-Have
Separate required knowledge from preferred qualifications. Required qualifications represent genuine minimum requirements, the skills the person must have to perform essential functions of the role. If you would interview someone without a particular qualification, it would be preferred.
For technical roles, be specific about:
Languages and frameworks (Python, PyTorch, Kubernetes)
Data and infrastructure (Snowflake, Databricks, AWS Vertex AI)
Domain experience (recommendation systems, NLP, computer vision)
Scope and scale (production systems, team leadership, cross-functional collaboration)
Step 4: Calibrate Seniority and Leveling
Ensure the job title, scope, and compensation align. A “Senior” ML Engineer has supervisory responsibilities or technical leadership scope that differentiates them from mid-level. Reference your internal leveling rubrics and market data.
Step 5: Confirm Compensation and Reporting Lines
Lock in the compensation range, equity package, and reporting structure before publishing. Misalignment here causes offer rejections and wastes everyone’s time.
How AI Tools Can Help
Fonzi’s multi-agent AI turns rough bullet notes from a hiring manager conversation into a structured draft position description. The system detects ambiguities, such as “optimize performance” without a metric, flags potentially biased language, and aligns responsibilities with market expectations for similar roles. This reduces drafting time from hours to minutes while keeping the hiring manager in final authority over the document.
Position Description Example: Senior Machine Learning Engineer (2026)
Here’s a concrete example for a Senior Machine Learning Engineer role at a fictional tech scale-up in 2026. Use this as a template you can adapt for similar AI and engineering positions.
Position Title: Senior Machine Learning Engineer – Recommendation Systems
Location: San Francisco, CA (Hybrid – 3 days in office)
Reports To: Director of Machine Learning
Team: ML Platform & Personalization (8 engineers, 2 data scientists)
Compensation Range: $210,000 – $260,000 base + equity
Position Summary
We’re looking for a Senior Machine Learning Engineer to own the end-to-end ML lifecycle for our personalized content ranking system. You’ll design, build, and deploy production models that directly impact engagement for 15M+ monthly active users. This role sits at the intersection of ML engineering and product impact. You'll work closely with product, data engineering, and infrastructure to ship models that move business metrics. Success in this role means shipping models that measurably improve user engagement while maintaining system reliability and scalability.
Key Responsibilities
Design and deploy ranking models to improve click-through rates by 15% by Q2 2027
Own model training infrastructure, including feature pipelines, experiment tracking, and model registry
Collaborate with data engineering to build and maintain feature stores on Snowflake and Databricks
Develop monitoring and alerting systems to detect model drift and performance degradation
Reduce model inference latency by 30% through optimization and infrastructure improvements by Q4 2026
Partner with product managers to define success metrics and translate business goals into ML problems
Mentor junior engineers on ML best practices, code quality, and production systems thinking
Participate in on-call rotation for ML infrastructure (approximately 1 week per quarter)
Required Qualifications
5+ years of experience building and deploying production ML systems at scale
Strong Python proficiency with experience in PyTorch or TensorFlow
Experience with recommendation systems, ranking models, or personalization at scale
Hands-on experience with MLOps tools (MLflow, Kubeflow, or similar)
Demonstrated ability to collaborate cross-functionally with product and engineering teams
Bachelor’s degree in Computer Science, Machine Learning, or related field (or equivalent experience)
Preferred Qualifications
Experience with transformers and retrieval-augmented generation (RAG) systems
Familiarity with real-time feature serving and low-latency inference
Experience with A/B testing and causal inference methodologies
Prior work with Snowflake, Databricks, or similar data platforms
Master’s degree or PhD in Machine Learning or related field
Experience leading technical projects or mentoring engineers
Tech Stack
Python, PyTorch, Snowflake, Databricks, Kubernetes, AWS (SageMaker, EC2, S3), MLflow, Airflow, Redis
Working Model
Hybrid – 3 days per week in our San Francisco office (Tuesday, Wednesday, Thursday). Must be able to overlap with PST business hours for real-time collaboration.
This example demonstrates how specificity helps job applicants self-select and helps interviewers evaluate against consistent criteria. Note how each responsibility includes measurable outcomes, and qualifications clearly distinguish between required and preferred.
Where AI Fits In: Fonzi’s Multi-Agent Approach to Position Descriptions
Traditional hiring teams struggle to keep up with role complexity and candidate volume, especially when hiring multiple AI and engineering roles in parallel.
Fonzi’s multi-agent AI reads and structures position descriptions into machine-readable components while flagging missing or ambiguous details for recruiters to refine. Instead of manually parsing hundreds of resumes against a wall of text, Fonzi’s agents extract the signals that matter and match candidates based on verified skills and experience.
This structured approach powers three critical capabilities:
Accurate Matching: By parsing position descriptions into discrete requirements, Fonzi matches candidates based on actual qualifications rather than keyword coincidence
Fraud Detection: Multi-agent analysis identifies potential misrepresentation in resumes and profiles, a growing concern in AI/engineering hiring where applications show signs of fraud
Structured Interview Guides: Fonzi generates interview guides tailored to the specific role, helping teams evaluate candidates consistently
Critically, hiring managers retain decision control throughout the process. Fonzi’s AI agents handle repetitive analysis and triage, reviewing resumes, flagging mismatches, and surfacing top candidates. Final candidate selection, interview decisions, and offers remain human-driven. You exercise final authority while the AI handles the workload of reviewing 250 applications.
Conclusion
Position descriptions are not just HR paperwork. They form the foundation of every successful technical hire. Clear responsibilities, outcomes, and requirements reduce friction across the hiring process. Interviewers align faster. Candidates self-select more accurately. Offers get accepted because expectations match reality.
For AI and engineering roles, specificity about tech stack, systems ownership, and success metrics is essential. A well-structured position description often turns a 90-day hiring cycle into 45 days by giving recruiters, hiring managers, and candidates a clear understanding of success.
Book a demo to see Fonzi in action, your next AI or engineering hire starts with a better position description.
FAQ
What is a position description and how is it different from a job description?
What should be included in an employee position description?
Can you show me an example of a well-written position description?
Who is responsible for writing and maintaining position descriptions?
How detailed should a position description be for internal vs. external use?
