Intrepid · Interview · Tue 10am 1 / 10
Tuesday 15 September · 10:00am · Level 7, 567 Collins Street

The roadmap
already exists.
You are not being
asked to write one.

The position description says the role will “deliver a predefined operational improvement roadmap.” Predefined. Rebecca and David have already decided what needs doing. They are hiring someone to carry it, not to redesign it.

The single most useful thing you can do tomorrow Ask to be walked through the roadmap as it stands, and which part of it worries them most. Then show them you would deliver it without losing what made it right.
Swipe across, or use the tabs above →
Section 02

Two people, two different questions

They want different things from you, and the answer that satisfies one can bore the other. Knowing which question you are being asked at any moment is most of the skill tomorrow.

Rebecca Conlin · Global Image Manager

The incumbent, going on maternity leave. Her background is production, not systems: Producer at WOW Pictures across fashion, corporate and food, and Content Producer at Daylesford Macedon Tourism before that.

The detail worth noticing in her title

She is Global Image Manager. The stills library is very likely the one that got properly structured, and video is the side that grew faster than anyone could organise it. That would explain why this role is scoped to video only.

So the generous question, and a genuinely useful one, is: “How much of what already works on the image side can carry across to video, and where does it break down?” It treats her existing work as the model rather than the problem.

  • Her real question is: can I hand this over and actually switch off?
  • What reassures her: curiosity about why things are the way they are, and respect for the parts that already work
  • What worries her: someone who arrives certain, changes everything, and leaves her a system she does not recognise in twelve months

David Nagle · Head of Global Brand Production

The hiring manager, and the person you would report to. Irish, History degree from Trinity, came into Intrepid from freelance, with a background across content management, online journalism, social and television. He leads global brand production.

The thing to know about his year

David’s team produced “Only Intrepid”, their largest brand campaign ever: close to 1,000 creative assets, a reported AU$5 million+ initial investment, built with Melbourne studio SouthSouthWest, running across digital, social, email, out-of-home, cinema, TV, print, podcasts and radio, with outdoor in Melbourne, Sydney, Brisbane, London, Toronto and New York.

One campaign, a thousand assets. That is the number that turns a video library from a folder problem into a findability problem, and it came out of his department.

  • His real question is: does this make my team faster, and will people actually use it?
  • What impresses him: production outcomes. Less rework, cleaner handovers, editors waiting on fewer things
  • What loses him: being out-teched. He is a producer and a content person, so translate every system answer into what it does for a shoot or an edit
Also in the building

Charlotte Day, Recruitment Partner APAC, or Andrew, will collect you from the Level 7 lobby. Charlotte has been warm and quick with you throughout, so say hello properly if you see her.

Section 03

Five accountabilities, and how each is scored

The PD lists five areas with a benchmark measure attached to each. This is effectively your twelve-month scorecard, written down in advance. Being able to name all five unprompted is the single clearest signal that you read it properly.

1 · Video Asset Management & Governance

Audit the iconik library. Design an improved folder structure and asset hierarchy. Define metadata, taxonomy and tagging specifications. Establish governance documentation for future uploads. Improve searchability and self-service.

Scored on
  1. Audit completed, new architecture implemented
  2. Specifications and metadata standards documented
  3. Searchability and self-service significantly improved
  4. Governance documentation adopted by team. Adopted, not written

2 · Business Requirements & Stakeholder Partnership

Identify and translate business requirements into workflow or system improvements. Partner with post-production and creative stakeholders. Communicate progress, options and recommendations.

Scored on
  1. Requirements documented and validated with stakeholders
  2. Solutions align with identified needs
  3. Positive stakeholder feedback. A soft measure, so it needs evidence you deliberately collect

3 · Video Workflow Optimisation

Improve post-production workflows including editing and proxy workflows. Standardise production handover. Improve collaboration between the editing team and wider stakeholders.

Scored on
  1. New workflows documented
  2. Reduced manual touchpoints
  3. Improved production consistency
  4. Reduced rework. This is the word to pick up and use back at them

4 · Improvement and Innovation

Review video software and subscriptions. Assess duplication. Recommend consolidation. Identify automation or AI-assisted tagging opportunities. Champion continuous improvement.

Scored on
  1. Tools reviewed and consolidated where applicable
  2. Cost saving opportunities identified. The only pillar with money attached
  3. Recommendations explored and documented

5 · Documentation and Backup

SOPs for post-production and DAM processes. Quick Start Guides and training documentation. Workflow charts. Governance, metadata and asset backup and retention standards. FAQ and knowledge base content.

Scored on
  1. Documentation library completed
  2. Increased self-service and reduced repetitive support requests
  3. Faster onboarding, consistent ways of working
The three phrases in the PD that most candidates will read past

“Predefined operational improvement roadmap.” The plan exists. Ask about it rather than replacing it.

“Rather than hands-on post-production execution.” You are not being hired to edit. Your shooting and editing is context that makes you credible, and it is not the job. Mention it once, then stay in operations.

“Project management systems such as Asana.” Named explicitly in the essentials. If you have used it, say so plainly. If not, say which you have used and move on.

Section 04

What is actually current, as of this month

You do not need all of this. You need two or three pieces used at the right moment, because specificity is what separates someone who has run a DAM from someone who has read about one.

Pillar 1 · iconik is mid-transition, right now

  • At IBC this month, Backlight announced AI-powered Discovery across iconik: AI-suggested metadata that editors review and approve, and search in plain everyday language instead of guessing exact tags. It builds on the Creative Operations Platform they introduced at NAB in April
  • Historically iconik’s native AI has covered speech-to-text transcription and basic object, face and logo detection. Semantic scene understanding, the “find me someone cooking at a homestay at dusk” kind of query, has been the gap. For a travel brand that gap is the whole ballgame, because travel footage is scenes rather than products
  • The Iconik Publish Panel is now generally available, bringing Wildmoka publishing into iconik so approved content goes to YouTube and Instagram with its metadata and approvals still attached
  • Worth knowing structurally: iconik indexes rather than stores. It sits over S3, on-prem NAS or Google Drive, which means storage and archive strategy is a separate decision from DAM strategy. That matters for the backup and retention accountability in pillar 5
The question this earns you

“Are you on the AI Discovery rollout yet, or still on the older transcription and object detection?”

Almost nobody will ask that. It also quietly says something true and generous: before building anything custom, check whether the platform has already shipped it. At Quad Lock you built the human-in-the-loop review console by hand. Backlight is now shipping that pattern natively.

Pillar 1 · the taxonomy call for a travel brand

Facets, not folders. A clip of a cooking class in Hanoi is Vietnam, and food, and small group, and a specific trip code, and 2026 campaign, all at once. A folder makes you pick one home for it. Faceted metadata lets it live in all of them, and iconik supports unlimited custom fields grouped into views for exactly this.

Folder structure still matters for ingest discipline and permissions. It stops being the way people find things.

Pillar 3 · proxies, and the distinction people get wrong

  • Review proxies and edit proxies are different jobs. iconik auto-generates H.264 720p or 1080p proxies for browser playback. That serves review and search. It does not serve an editor who needs to relink to camera originals
  • A relinkable proxy workflow needs matching clip name, timecode and record trigger, so the proxy can relink to the original in Premiere or Resolve. Get that wrong and you create the rework the PD is asking you to reduce
  • Camera to cloud puts lightweight proxies into an edit timeline while a shoot is still running. For a brand shooting across 100+ countries, that means footage is reviewable before the crew has landed, and dailies reach stakeholders without a hard drive in a bag

Pillar 4 · where the cost saving actually sits

  • Seat audits first. iconik is per-user monthly, publicly quoted at roughly $50 to $150 per user, plus storage and transcode charged by the minute. Per-user pricing punishes dormant accounts, so inactive seats are the fastest saving in any DAM and the least controversial one to propose
  • Overlap is the second. Most brand teams pay separately for review and approval, large file transfer, transcription and publishing, and a DAM that has quietly grown all four. The Publish Panel going GA is a live example
  • Storage tiering is the third. Because iconik indexes rather than stores, cold archive can sit behind it and the metadata stays searchable even when the file is in a cheap tier

The frontier, if it comes up

Multimodal video embeddings are the thing that finally makes untagged footage searchable by meaning, indexing visual, audio, spoken word and on-screen text together rather than in separate pipelines. Four days ago TwelveLabs’ Marengo model became the first video model available in Amazon Bedrock’s managed knowledge base.

How to use this without overreaching

Mention it as a direction, not a proposal. “That is where this is going, but I would not touch it in year one. You cannot put semantic search on top of a library you have not audited yet, and the taxonomy work still has to happen first.”

Knowing when not to use the shiny thing is the senior signal here.

Section 05

You have already built four of the five

Take a minute with this before you go in. The intimidation you felt about the number came from the salary, not from the work, and the work is work you have done.

Pillar 1
Asset management & governance. You organised and tagged a 500,000-asset library by hand, then defined a classification-first schema so each asset carries only its relevant fields, then built a precision and recall gate that decided when tagging could write back unattended
Pillar 2
Stakeholder partnership. That library served marketing through to product design. The Slack search tool exists because you worked out what people actually needed rather than what the system offered
Pillar 3
Workflow optimisation. The thinner one on paper. Your strength here is that you shoot and edit, so you know where handovers break. Use it as credibility, not as the main claim, because the PD says this role is not hands-on post
Pillar 4
Improvement & innovation. You identified AI tagging as the answer and then architected it, which is the desirable line on their list, and you built the review console that made it safe to use
Pillar 5
Documentation. Runbooks and troubleshooting guides a non-developer maintains. That is the exact benchmark: self-service and reduced repetitive support requests
The sentence that ties it together

“I did this in the order your roadmap describes, except I did the manual half first. Hand-tagging half a million assets is what taught me which metadata schema was worth building, and I do not think I would have got the schema right without it.”

That is a better answer than any credential, because it explains why your judgement is trustworthy rather than asserting that it is.

Where you are genuinely lighter, and how to hold it

  • Asana. Named in the essentials. Say what you have actually used to run projects and move on. Do not oversell familiarity with a tool they will put you in on day one
  • Scale of stakeholder spread. Intrepid runs 35 country offices. Quad Lock was one company in one place. If it comes up: “the spread is new to me, the pattern is not, and I would want to know early which offices actually generate footage and which just consume it”
  • Years in post. Here is the thing I got wrong earlier and it is good news: the PD they sent you contains no years requirement at all. It asks only for “experience working within video production, post-production or creative operations.” The “5+ years” line was in the Seek ad and it was dropped from the formal position description. So there is nothing to defend. Do not raise it, and if it comes up, seven years from Ubisoft in 2019 through to now across production, post and creative operations is the honest count
Section 06

What they said they will ask

Charlotte wrote it out: they will dive deeper into your experience, how your skills and values align, “how you’ve approached challenges in your career, the skills you can bring to this role, and what excites you about this opportunity.” That is a behavioural interview. Have four stories ready, not four speeches.

Walk us through what you built at Quad Lock.

Chronological and concrete, because the order is the proof. Hand-tagging, then the schema, then the AI layer, then the accuracy gate, then Slack search, then runbooks. Keep it to ninety seconds and let them pull on whichever thread interests them.

Cover these
  1. 500,000 assets organised and tagged by hand first
  2. Classification-first schema, each asset carries only its relevant fields
  3. AI tagging with human-in-the-loop review, per-field accept, edit, bulk approve
  4. Precision and recall gate before anything wrote back unattended
  5. Plain-language search in Slack, plus runbooks a non-developer maintains
Tell us about a challenge you worked through.

Their pack says they look for people who “stay genuine when things get challenging”, so the point of this answer is temperament, not triumph. The Aluvers shoot works well: the components delivery got missed, you spotted mid-shoot that the film had a hole where the product introduction should sit, and you staged one shot the next morning to patch it. Small, real, and it shows you were editing in your head while still shooting.

How would you get people to actually adopt a new system?

This is the question underneath the whole role, because “governance documentation adopted by team” is a literal benchmark measure. The honest answer is the Slack tool.

The shape “I stopped trying to get people into the system. Nobody was going to learn a new interface to find one clip, so I put plain-language search into Slack where they already worked. Most adoption problems are design problems wearing a costume. And if a metadata field gets skipped every time, the field is wrong, or it is in the wrong place in the workflow. Mandatory at upload gets you rubbish. Mandatory at approval gets you real data, because someone is already looking.”
What would you do in the first few months?

Lead with the roadmap, not with your own plan. This is where most candidates will overreach.

The shape “First I would want to properly understand the roadmap that already exists, because you have both thought about this for longer than I have. Then I would watch how people actually look for footage rather than how the process says they should, since the gap between those two is usually where the real problem is. Then audit and get a number on it, so that what comes next is evidence rather than opinion. And I would write the documentation as I go rather than at the end, because it is twelve months and the thing has to work after I have gone.”
Cover these
  1. Their roadmap first. Understand it before adding to it
  2. Watch real search behaviour, do not survey it
  3. Audit and baseline. Time to find, duplication, rework
  4. Document as you go, because it has to outlive the contract
How do you handle usage rights and permissions?

Strong ground for you. Rights have to travel with the asset as structured fields rather than a PDF in a folder: channels, territory, term, editing rights, ownership, expiry. And it connects directly to the retention standards in pillar 5.

Cover these
  1. A signature someone did not understand is not really consent. You write releases in plain language for that reason
  2. Store the exact wording plus a timestamp, so you know what was agreed, not just that something was
  3. Whatever is not a field is a rumour. Rights in a PDF cannot be queried or flagged on expiry
  4. Be straight that you have not run a rights register at their scale, then say how you would start one
What excites you about this opportunity?

Keep it short and make one of the two reasons about them rather than about the work. The climate decision is still your strongest, and you have standing on it through Climate Fresk. Then the honest one: this is the first ad you have seen that wants both halves of what you do.

Which of our values resonates most?

Pick one, evidence it, stop. Reciting all four is the answer everyone gives. Real is the one you can actually evidence, because you name what you cannot do rather than overselling, and you have put honest gaps in writing on applications.

Salary expectations?

$90,000 to $98,000, said plainly. The ad’s ceiling is $98,000 and Melbourne medians for comparable scope sit slightly above it. Give the range, stop talking, let the silence sit.

Section 07

Charlotte said two-way conversation

She wrote it in the invitation, and their candidate pack says the same thing. With the incumbent in the room, your questions are worth more than your answers, because they are the only part of the hour where they see how you would actually start.

The one that matters most, and ask it early

To both of them The PD mentions a predefined improvement roadmap. Could you walk me through where that has landed, and which part of it worries you most?

It shows you read the document properly, it hands them the floor on work they are proud of, and their answer tells you what the job really is for the rest of the conversation.

For Rebecca

  • “When someone needs a clip today, what do they actually do?” The real behaviour, not the documented process
  • “How much of what works on the image side carries across to video, and where does it break?” Treats her existing work as the model
  • “What did you want to fix and never get the time for?” This is the actual brief, and it is rarely identical to the ad
  • “Is there anything you would want me to leave alone?” Generous, disarming, and it shows you understand what handing over feels like

For David

  • “After a campaign the size of Only Intrepid, where did the asset library feel the strain first?” Specific, flattering, and straight to why the role exists
  • “Twelve months from now, what has to be true for this to have worked?” Makes him define success out loud, in front of you
  • “Who owns the system after the contract ends?” Shows you are designing for handover, and he may not have answered it yet
  • “Are you on iconik’s new AI Discovery rollout, or still on transcription and object detection?” Only if the conversation is already technical. It is the most expert question on this page
  • “What has been tried before that did not stick?” Most organisations have a failed DAM attempt in the cupboard

One human one, near the end

  • “What surprised you about working here?” Their pack invites exactly this, and it gives you something real to follow up on afterwards
Pick four, not thirteen

The roadmap question, one for Rebecca, one for David, one human. Listen to the answers rather than working down the list. Let one turn into a proper back and forth, because that is the bit they will remember.

Section 08

On building the mini-sites tonight

You told Cam you were thinking of building deployed mini-sites showing new workflows, audits, documentation and automations for Intrepid. The instinct is right. I think the timing is wrong, and I want to be straight with you about why.

Three reasons to hold it

  1. The roadmap is predefined. The PD says so in the position purpose. Arriving with your own roadmap for their library, built overnight without having seen it, risks reading as not having listened to the one sentence that matters most in the document
  2. Rebecca built the current system. A polished external plan for her library, presented to her while she is handing it over, is a hard thing to receive gracefully however kindly it is meant
  3. Nobody asked for a presentation. Charlotte described a two-way conversation about your experience, your challenges and your values. Turning that into a pitch changes the format on them, and the format was chosen deliberately
The version I would actually do

Bring proof of what you have already built, not a plan for what they should build.

A short case study of the Quad Lock system, mapped against their five accountabilities. It is unfalsifiable, it cannot land as criticism, and it answers the only real question they have about you: has this person done it before.

Offer it rather than present it. “I put together a short thing on how the Quad Lock build maps to your five areas, if that is useful. Happy to send it after.” Then let them ask.

And hold the real build for round two

Their process runs at least two interviews. Round two is where a deployed prototype becomes extraordinary, because by then you will have heard the roadmap and can build against what they actually said rather than what you guessed.

Same effort, a fraction of the risk, and ten times the accuracy. If you get through tomorrow, we build it properly with real information.

On Cam

He said he would come back to you. If he replies before tomorrow morning, his read beats mine on anything about how you actually worked, because he watched you do it. Send me whatever he says and I will fold it in.

Section 09

Intrepid, properly

Where it came from

In 1988 two Australian university friends, Darrell Wade and Geoff “Manch” Manchester, drove a second-hand truck from London to Nairobi over nine months with twelve other people. They came back convinced there was a better way to travel than a coach and a hotel.

Their first trip, in 1989, was Thailand with five travellers and one tour leader: Geoff himself. It is now the largest adventure travel company in the world. Both founders are still on the board, Darrell as Chair.

Why that matters to you

A second-hand truck and five customers is a real starting line, and you have your own version of it. You started a studio from nothing in 2023. You do not need to labour the parallel, but it is honest, and it is a warmer answer to “why us” than admiring their size.

The numbers worth knowing

Scale
900+ trips, 100+ countries, run through their own network of 35 country offices rather than subcontracted out. 3,500+ staff
Revenue
$626 million in 2024, their record year
B Corp
Certified 2018. Recertified early 2025 at 102.5, up from 91.2 in 2021Only 20% of B Corps with 1,000+ employees score above 100
The Foundation
AUD $20 million+ raised for 160 charities since 2002
Their own people
80% of staff say the purpose work is a key reason they stayFrom their own engagement survey. A very good number to have read
Values
Real · Together · Ambitious · Impactful. They say plainly they assess against these

Who runs it

  • James Thornton, CEO. Darrell Wade, co-founder and Chair. Geoff Manchester, co-founder and Director
  • Hazel McGuire, Chief Marketing Officer, owns brand, PR, communications and customer experience. Your work would eventually serve her side of the house
  • Meegan Marshall, Chief People and Purpose Officer. Anu Karunatilaka, CTO, runs digital and data platforms
The pattern, and this is the thing to actually say

Almost every candidate will name one good thing Intrepid did. The stronger move is to name the pattern, because it shows you looked at more than one page.

Intrepid repeatedly gives up something valuable to do the right thing, before anyone makes them.

  • They were the first global travel company to ban elephant rides, roughly a decade ago, with World Animal Protection. That was revenue they walked away from while it was still normal in the industry
  • In October 2025 they retired their carbon offset program and gave up Climate Active certification, losing the right to call themselves carbon neutral, and put $2 million a year into a Climate Impact Fund for actual decarbonisation instead. Electric vehicles for suppliers in India and Nepal, solar on Cambodian homestays. New target: 8% Scope 3 intensity reduction per customer per day by 2030
  • They then open-sourced their B Corp playbook so competitors could copy it

Giving up a marketing claim is harder than earning one. That is the sentence.

And this is where you have standing

You facilitated Climate Fresk workshops for two years, walking groups through exactly the gap between offsetting and reducing. So you know how easy the other choice would have been. You have earned an opinion here rather than borrowed one, and that difference is audible.

What they are doing right now

  • Their largest ever product launch: 100+ new trips for 2026 across 50+ countries, with two new ranges, Short Breaks and Premium Walking and Trekking
  • 28 new active trips taking the active range to 148, on the back of 9% growth in active bookings
  • 10 new family trips taking that range to 50, plus 20% off for travellers 17 and under, after 15% growth in family travel
  • Women’s Expeditions expanding into Peru, Cambodia and Bhutan in 2026, on top of India, Morocco, Nepal, Saudi Arabia and Turkey. Locally led, and they exist partly to create employment for local women
Connect that to the actual job

A hundred new trips across fifty countries is an enormous amount of new footage arriving at once, from country offices all over the world, in every format and naming convention imaginable. That is precisely the problem this role exists to solve, and it is happening right now rather than hypothetically.

Saying that out loud shows you understand why they are hiring, which is a different thing from understanding the job description.

Three things to avoid saying

  • Do not recite the four values back at them. Pick one, evidence it, stop
  • Do not gush about the trips generically. Everyone does. If you have not travelled with them, do not imply you have
  • Do not describe them as a tour company. They are a B Corp that runs its own operations in 35 countries, and the distinction is the thing they are proudest of
Section 10

Tuesday morning

Arrive
9:55am. They asked for five minutes early, so five minutes early is the right answer rather than twenty
The route in
567 Collins Street. Lifts on the right hand side, select Intrepid on the touch screen for Level 7, it directs you to a lift. Turn left into the lobby, ask for Andrew or Charlotte
Dress
No formal dress code, in their words. Smart casual, something you feel like yourself in
Stage
Their process runs at least two interviews. Tomorrow is not the decision. Aim to be someone they want to see again

Take with you

  • Four questions written down. Looking at a note to ask a question reads as prepared, never as unsure
  • Your phone with the portfolio already open on the right page, in case the conversation turns to the work
  • One printed resume in your bag. You will probably never take it out
  • Water, and eat something first. It is a 10am and you will talk more than you expect

The four things to leave in the room

  1. You have built this before. iconik, taxonomy, AI tagging, accuracy gate, runbooks, in that order
  2. You did the manual half first, and that is why your judgement on schema is worth something
  3. You would deliver their roadmap, not replace it
  4. You design for handover from week one, because it has to work after you have gone
The whole morning in one line

Curious first, competent second, warm throughout. The competence is already established or you would not be in the building. The curiosity is what makes someone want to hand you their work.

Afterwards

Short note to Charlotte.Day@intrepidtravel.com that afternoon. Three lines, one specific thing from the conversation, no pitch. Name Rebecca and David, because that is the part that gets passed on.

Last thing

You shot with Pluto this morning on a Netflix comedy, you have a partner you are happy with, and it is spring. You are walking in from a good place rather than a desperate one, and that is audible in a room. Let it be.

Mounir ProductionsInterview brief · Rebecca Conlin & David Nagle
Intrepid TravelVideo Operations Implementation Lead