COACH’s One Internship Puts Demonstrated Skill First

The Hiring News: COACH Is Looking for One Intern

COACH, the San Francisco AI job-search platform operating as HiringCoachAI, has one open position on its careers page: an AI Automation Intern (Summer 2026), remote or based in San Francisco. The careers page at hiringcoach.ai/careers carries a single message: "We're a small team building the tools we wish we'd had during our own job searches. If that resonates, we'd love to hear from you." The company's LinkedIn page lists 2–10 employees and 1,922 followers.

The platform has been featured twice at United Nations Career Week, at the Peace Corps Career Fair, at the State Department Career Fair, and was selected for the NVIDIA Inception Program. User testimonials on the site describe the tools as "a brilliant mentor who understands the language of development, equity, and global action" and credit the platform with translating public-sector experience into private-sector offers.

HiringCoachAI is a job search platform for job seekers — not an employer hiring coaches. It helps users tailor resumes and cover letters to specific roles, check fit before applying, practice interviews, draft outreach messages, and track applications in one workspace. The platform is built for people the system overlooks: accomplished public sector, nonprofit, and international development professionals making a career change.

How AI Screens Work Across the Industry

AI-powered candidate screening is now standard. Eighty-seven percent of companies used AI in recruitment in 2026, with over 65 percent of recruiters relying on it for hiring decisions.

These platforms share a common architecture. Sapia.ai describes its approach as "structured chat" that "helps you assess everyone consistently, identify stronger candidates and give personalised feedback at scale." Hirevire frames the shift as "moving beyond traditional resume screening to objective, skills-based evaluation" using AI to "assess candidate capabilities, reduce unconscious bias, and improve hiring quality while accelerating the screening process." Gomokka defines the tool class broadly: software that uses artificial intelligence to evaluate job applicants' skills, competencies, cognitive traits, or job fit, ranging from resume screening algorithms and skills tests to AI-conducted pre-interviews and job simulations. Talentprise summarizes the method: "AI candidate screening matches candidates by skill and context, not keywords." Hiremore.ai breaks the criteria into two types — structured rules and weighted signals — that "filter out unqualified candidates and ranking those who remain." Imocha's survey of the top 15 tools highlights features like skills testing, AI proctoring, candidate scoring, and advanced analytics.

The research suggests a typical screen operates on three layers: a parsing layer that ingests resumes and work samples into standardized skill tags; a scoring layer that weights those tags against role-specific criteria (required qualifications, preferred experience, must-have skills, nice-to-have attributes); and a ranking layer that produces a shortlist for human review, often with a fit score and evidence-gap summary. The ICF AI Coaching Standards framework adds a governance overlay: expectations for transparency, bias mitigation, and data privacy in any AI system that evaluates coaching competence.

What the research does not show is COACH's specific criteria weights, threshold scores, or whether it uses a vendor or proprietary model for its own hiring. The HiringCoach.ai platform (which job seekers might use to prepare for such screens) offers fit analysis against a posting, tailored resume bullets, cover letter drafts, and verbal interview practice with scored feedback, but it does not sit on the employer side of the table.

What the Intern Role Demands

The single posted role is the internship. The careers page does not publish detailed specifications. Based on the platform's described functionality (AI-driven resume tailoring, cover letter generation, interview prep, job board search across 10+ sources, and a "Mission Control" workspace that prioritizes next steps), the internship likely involves work on the automation pipelines that power those features.

The platform's own description: "the same message." The intern would join a team of 2–10 employees in San Francisco (or remote) building AI job-search tools for career switchers and public-sector professionals.

Candidate Strategies: How Applicants Are Getting Past AI Screens

Candidates advancing past AI screens share a pattern: they treat the screening layer as a distinct technical challenge, not a formality. Research across AI hiring platforms shows the screen evaluates resumes and recorded responses against weighted criteria pulled from the job posting: scoring structure, keyword alignment, and evidence of specific capability before any human review occurs. If the resume does not clear that filter, the application dies silently with no rejection email and no explanation.

Resume Engineering for Parsing and Scoring

Successful applicants start with format discipline. A clean single-column layout using standard headings (Work Experience, Skills, Education, Certifications) remains the safest choice because parsing software is trained to find those exact labels. Creative headings like "My Journey" or "What I Bring" look distinctive to humans but confuse the parser. Graphics, tables, columns, icons, and text boxes are invisible to text extractors and often cause critical details to be dropped. File type matters: PDF or .docx parse reliably; image-based PDFs, .pages, and .odt files frequently fail. Standard fonts (Arial, Calibri, Georgia, Times New Roman) render consistently; decorative fonts are stripped or misread. Headers and footers are ignored entirely: contact information belongs in the document body. Consistent date formatting (e.g., "Jan 2023" throughout) prevents timeline parsing errors that can scramble experience chronology.

Keyword alignment comes from the job description itself. Candidates extract 8 to 10 role-specific terms (tools, skills, certifications, platforms) and mirror the exact phrasing wherever it honestly matches their background. If the posting says "Digital Marketer," writing "Marketing Specialist" reduces match scores. Using "CRM" when the description uses "Customer Relationship Management" (or vice versa) creates unnecessary friction. The goal is not stuffing; it is natural placement in skills and experience sections where the terms reflect real work.

Achievement bullets replace duty lists. "Responsible for reports" scores lower than "Created weekly Excel reports, reducing manual reporting time by 30%." Numbers outperform adjectives in automated scoring: "Completed 300+ annotation tasks per week at 98% accuracy" beats "Managed a high-volume annotation workload." For technical roles, leading with domain expertise and explicit language proficiency ("Native Arabic speaker," "C2 Spanish") and prior AI-task experience ("Completed 200+ RLHF evaluation tasks on [platform]") add scoreable signals that generic resumes lack.

Video Interview Tactics: Structure, Timing, and Authenticity

When the screen advances to recorded responses, the first 30 seconds determine whether a human recruiter samples the clip. Candidates who open with the conclusion — "In that situation, I reduced processing time by 60% by building an automated triage system (here is how I got there)" — outscore those who lead with situation setup. The STAR format (Situation, Task, Action, Result) structures answers the algorithm recognizes, but the result must be quantified and the answer must close on that outcome, not trail into a lesson learned.

Answer length has a sweet spot: 90 to 150 seconds consistently scores higher than responses under 60 seconds (too thin) or over 180 seconds (unfocused). Practicing aloud, recording at least three full answers, then reviewing for filler words, incomplete sentences, and missing outcomes, closes the gap between imagined fluency and delivered performance. The platform's practice mode is a technical setup test for camera, microphone, and lighting; candidates who treat it as prep waste the only risk-free diagnostic.

Clear, steady speech reduces transcription errors that degrade content relevance scores. Rapid delivery or heavy filler ("um," "like," "you know") produces a worse transcript and a lower score. Reading from a script backfires: eye-tracking detects the zigzag pattern, and anti-cheating mechanisms flag keyword stuffing or AI-generated response patterns: identical phrasing, identical example sequences, identical namedropping across candidates. Deepfake substitution is caught by facial recognition tools now standard on major platforms.

Differentiation Over Conformity

A career coach cited in Business Insider said AI tools compare talking points across the candidate pool and rate them accordingly. The first few points that come to mind are what 99 percent of candidates will say. Advancing requires pausing to identify what the average applicant will offer, then selecting a distinct, verifiable example. Professional presentation still matters: overdressed beats underdressed, and a neat, tidy background signals the same judgment a human sampler makes in seconds.

The retake button is a trap for minor nerves. Retakes reset the entire attempt; candidates who re-record for a stray "um" often perform worse under added pressure. Save it for factual errors or running out of time before the close.

Pre-Submit Validation

Tools like RoastMyCV (built into Talents for AI) analyze a resume against real recruiter criteria and AI screening standards before submission. Upload the CV, enter the target role and location, and the tool flags sections most likely to be filtered out. Cross-checking LinkedIn and job-portal profiles for consistency with the resume prevents mismatches that trigger parser confusion.

The through-line across every tactic: the screen rewards specificity, structure, and honesty. Candidates who optimize for the parser's actual logic — not their guess of what "looks good" — are the ones who reach human review.

The Bigger Picture: AI Coaching Platforms and the Shifting Hiring Landscape

The AI coaching market has moved past the hype cycle. In 2026 it is a functioning product category with paying clients, measurable retention, and pricing that undercuts the lower end of human coaching.

Adoption signals are concrete. Fifty-two percent of coaches now use AI tools, up from 15 percent in 2020. Seventy-one percent of enterprise companies had adopted AI coaching platforms by Q4 2023. Seventy-six percent of corporate HR departments were piloting AI coaching programs in 2023. About one-third of Fortune 500 companies now rely on executive coaching to strengthen leadership pipelines.

The competitive landscape has sorted into three tiers. Pure-play AI platforms (Hupo, Hyperbound, Rocky.ai, and a wave of job-search automation tools like JobHire.AI and LazyApply) deliver continuous, personalized feedback at a scale no individual human coach can match. Hybrid platforms (BetterUp, CoachHub, Coachello) pair ICF-certified coaches with AI roleplay, practice environments, and workflow integration inside Slack and Microsoft Teams. Coachello reports managers on its platform log 3.2 coaching interactions per week versus 0.7 for traditional programs. BetterUp publishes a 56 percent improvement in resilience and 26 percent decrease in stress over six months. CoachHub's AIMY matching drove a 41 percent improvement in manager effectiveness scores over 12 months at a global manufacturer. Human-only platforms (BetterUp's core offering, CoachHub's human tier, EZRA, Torch) operate at €3,000–€8,000 per person per year, serving the executive and high-stakes segment where relational depth and nuanced judgment remain the differentiator.

The market is bifurcating. Competitive pressure from AI platforms concentrates at the entry level, specifically below $150–$200 per month for online coaching. Coaches whose programs sit in that range without a clear differentiation narrative are the most exposed. The minimum defensible price for human coaching in 2026 is somewhere in the $200–$250 per month range for online-only delivery, and only if the program structure justifies it. What AI platforms cannot credibly replicate — reading what a client doesn't say, nuanced program adjustment based on psychological context, accountability with real social stakes, credible outcome guarantees with teeth, complex technique correction — defines the human premium.

Hiring is moving in parallel. Tools like HiringCoachAI now cover the full coaching workflow: value proposition, role research, resume tailoring, interview prep, offer negotiation, daily next steps.

COACH sits inside this shift as a job-search platform for the overlooked. Its one open internship and the AI-powered screens that candidates will face reflect the same logic driving the broader market: scale, consistency, and evidence-based evaluation.

What's Out of Scope: What This Article Does Not Cover

This article examines COACH's single open internship and the AI-driven screening criteria that determine who advances to human review across the industry. That scope is deliberate. Several adjacent topics, important in their own right, fall outside it.

Other AI coaching platforms' hiring practices are not analyzed here. Each company designs its own evaluation pipeline. Some weight credentials heavily; others prioritize demonstrated ability. Comparing those pipelines would require a separate piece with direct access to each platform's screening logic.

Ethical debates surrounding AI in coaching are also excluded. The research literature flags substantial concerns: data bias that can compromise training outcomes and create unjust resource distribution; vulnerabilities in data storage that risk unauthorized leakage of sensitive information; ambiguity regarding accountability when AI-generated guidance causes harm; and the broader question of whether AI should supplement or replace human coaches in high-stakes contexts. Frontiers in Digital Health (2026) documents these issues extensively, including calls for data minimization, robust encryption, transparency requirements, diversity in data collection, bias detection mechanisms, continuous monitoring, explicit role definitions, transparent accountability mechanisms, compensation frameworks, and mandatory human approval of AI-developed training plans before implementation. Those discussions matter — they shape regulation and product roadmaps — but they do not change what industry screens evaluate today.

General job market conditions are not addressed. The coaching industry is valued at $25 billion, and enterprise platforms typically price coaching at $3,000–5,000 per user per year, limiting access to senior leaders. Macroeconomic trends, funding cycles, and competitor hiring freezes all affect candidate supply. This article does not model them.

Broader AI hiring trends across tech are similarly out of scope. Companies such as Databricks and Anthropic operate different screens for different functions. Their salary bands reflect distinct talent markets. COACH's criteria are specific to its internship; they are not a proxy for how AI labs or data platforms evaluate engineers.

The article also does not cover certification frameworks, academic competency models, or international adaptation of coaching standards. Research on professional tennis coaches in China identifies five dimensions (Situational Judgment, Professional Integration, Digital-Oriented Coordination, Growth Orientation, and Humanistic Guidance) and recommends practice-based evaluations over theoretical examinations. The OECD defines core competency as an integrated construct of key competencies and professional qualities for complex situations. These frameworks inform the field but are not the lens COACH's screen applies.

Finally, this piece does not evaluate COACH's product efficacy, customer outcomes, or competitive positioning. Boon reports 23 percent average competency improvement and 89 percent session attendance across 110+ enterprise customers. Risely reports 26 percent average skill improvement in 12 weeks across 5,000+ users. Human-plus-AI coaching yielded roughly 74 percent higher weight loss than AI alone in a Stanford study. Those metrics belong in a product review or market analysis, not a hiring guide.

What remains is narrow and actionable: one internship, the industry-standard screen candidates will face, and the evidence that moves a candidate past it. The screen reads what you show. The rest is noise.

Working in AI? Zero G Talent tracks the openings: see every open Databricks role, browse AI jobs, openings at Anthropic, and the people building the field.