Roles that pay for what you actually do: framing the problem, designing the system, shipping it end to end, and owning the commercial outcome. No coding required.
Swipe right to keep, left to bin, down to park for later. Arrow keys work too.
Aggregators, the contract market, recruiters and the routes that never appear on a board at all.
Labs and AI-native startups hire off their own boards and through VC talent networks. By the time a role reaches Indeed it is usually well into its shortlist.
The industry runs on its own boards. Roles here rarely reach Indeed, and when they do it is late.
Filtered to the last 7 days where the site allows it.
AI companies and sports tech firms post to Greenhouse, Lever, Ashby and Workable days before the aggregators pick it up. These go straight at the source.
Total pay in London unless stated. The bar is the typical spread, the gold line is the median. Sanity check an offer with these, do not treat them as a promise.
The red line on every bar is your 60k floor. Anything to its left is not worth the move.
Four reads. One: AI transformation and enablement pays roughly double e-commerce management for work closer to what you already do. Two: in sport the vendors pay properly and the clubs do not, so aim at the companies selling into football rather than the clubs themselves. Three: your 60k floor rules out two whole families. E-commerce Manager sits almost entirely below it, and so do UK Sport and general club-side roles, which is a real cost of the football route rather than a technicality. Four: the frontier labs are the highest ceiling on this board by a distance, and the eye-watering numbers reported in London are engineering and research packages. The commercial and enterprise roles are the ones open to you, and at Anthropic's London office those are understood to approach 200k.
The top 1% of UK earners starts at roughly 207k. 300k is inside the top 0.5%, and around 300,000 people in the country clear it. It is a real target, but it is not a promotion away, and no ladder of ordinary salaried jobs reaches it in five years from where you are.
The structural point that decides everything else: above about 150k, UK income stops being salary. Partnership and dividend income make up more than a quarter of top-1% income and a third at the top 0.1%. People at 300k are not mostly people with a 300k salary. They are people with a 120k salary plus something that scales: commission, equity, or their own book of business. Pick which of those three you are building, because they need different next moves.
Run the stack, not a single job. Your position is unusual in that you already have two engines: a salary and a consultancy. Most people chasing 300k have one.
The honest probability. Reaching 300k in five years as a salaried employee alone is unlikely, maybe one in ten. Reaching it as salary plus a real second income is perhaps one in three. Reaching 180k to 220k in five years is very achievable and would put you in the top 1%. Plan for that, and treat 300k as the outcome if the equity or the business lands rather than the thing you are counting on.
What will not get you there: staying in e-commerce management, chasing job titles rather than compensation structures, or waiting for a big employer to promote you into it.
1. Sky is missing. The CV has Kub Products as 2023 to Present. You have been Trading Performance Manager at Sky Broadband since June 2026. Every AI adoption and enablement role on this board wants proof you can work change through a large organisation, and the one line that proves it is not on the page. This is the single highest-value fix and it takes ten minutes.
2. The CV and the board are chasing different jobs. The header says Head of eCommerce and Commercial Strategy. That is a good CV for the three e-commerce entries on this board and the wrong CV for the other eighty-six. There is no mention of AI, no Fulcrum, no agents, no MCP, nothing about building working systems with AI tooling. That is the thing that makes you unusual, and it is currently invisible.
What is already strong. The achievements are specific and numeric, which most CVs are not: profitability up 12 per cent on a 4m base, meta ads from 3k to 21k a month, ROAS 1.2 to 4.3, 20 hours a week freed through automation, 400k monthly impressions. Keep all of it. The problem is framing, not substance.
The reframe. Those automation and dashboard lines are currently filed as e-commerce housekeeping. They are actually the proof for the AI roles. Freeing 20 hours a week through automated workflows is an adoption case study; you just have to call it one.
Same evidence, different order and different language. Do not write a new CV each time; run the master through the relevant block.
What you sent me, unchanged, so tailoring always starts from the same place. Fix the two problems above and update this.
Aaron Manu · London · aaron.manu@btinternet.com · +44 7715 250594
Head of eCommerce and Commercial Strategy. Kub Products Ltd 2023 to present, Commercial Executive placement 2021 to 2022. Earlier: BT cybersecurity and sales intern, Cisco product and innovation intern, Warner Bros, retail operations. Loughborough BA Politics with Economics 2:1. Watford Grammar AAB. Shopify, Amazon, eBay, TikTok Shop, GA4, Clarity, Meta and Google Ads, Make.com, Xero, Looker.
Every card on the Board has a Tailor CV button. It copies a prompt carrying your master CV, that role's details, and the tailoring block for its family. Paste it into Claude and you get a version written for that posting.
Do not send a tailored CV without reading it. The tailoring reorders and relabels what you have done. If it ever claims something you did not do, that is on the page with your name on it, not mine.
How much to change. The profile paragraph, the top four achievements and the skills line, every time. The employment history stays fixed. If you find yourself rewriting the history, you are applying to the wrong role.
Ranked by how much each one is costing you against the families you are actually chasing, not by how hard it is to fix. Most of these are not skill gaps at all, which is the useful part.
Each one produces an artefact you can link to. Reordered as you swipe, so the top of this list is what serves the roles you are chasing today.
Only the ones that change which roles you can credibly apply to. The order follows your swipes: bin a family and its skills drop down the list.
What you should not do in those 90 days: a certification. See below.
The one certification exception: if a specific job description names a credential, get that one. The most valuable certification is the one that appears in the posting you are applying to. That is a reactive decision, not a plan.
The gap most people have is either commercial depth or the ability to actually ship something. You have both, plus large-company exposure at Sky and owner-operator exposure through Fulcrum. The risk is being read as "manager who is keen on AI." The fix is to lead with artefacts, not enthusiasm.
The labs and AI-native startups do not hire like normal tech employers, and applying to them like normal tech employers is why most people never hear back. OpenAI took over 400,000 applications in a single year.
More roles come from three moves, in this order of return.
What not to do: add more general job boards. Indeed, Reed and LinkedIn already cover that ground, and a fourth aggregator adds volume without adding signal.
Channels does not go stale, it queries the boards live. Board is a snapshot. The sweep now follows a written runbook, job-radar-refresh.md, which holds the candidate filter, the source order, the pruning rules and the standing questions. Paste this into a new chat:
/captain refresh my Job Radar. Read job-radar-refresh.md and job-radar-captain-log.md first, then run the sweep exactly as the runbook sets out. Keep my saved statuses and entry ids, flag what is new, and hand back the updated job-radar.html.
The Refresh button queues a real sweep. A webpage cannot wake Claude up, so the button does the next best thing: it drops a request in a queue. The agent on your Mac checks that queue every 15 minutes, runs the full sweep through Claude Code, edits this file, and redeploys it. Reload the tab twenty minutes later and the new roles are there. The same sweep runs on its own every day at 06:50.
No API key involved. The sweep runs on your Mac through Claude Code on your normal subscription. The queue is a postbox with no AI in it.
What breaks it: your Mac being off or asleep. The sweep runs locally, so nothing happens until it wakes. If a queued sweep has not landed after an hour, check sweep.log in the Job Radar folder.
Where your decisions are stored. Statuses and notes live in the browser you are using, on the device you are using. They do not sync between your laptop and your phone, and clearing site data clears them. If cross-device sync matters, say so and the next build can put a small Cloudflare Worker and KV store behind it.
Why the runbook matters. Without it every refresh restarts from instinct and the board drifts. With it the sweep works highest-yield sources first, dates every entry, refuses to promote listings it cannot date, and never renames an entry id, which is what your saved statuses are keyed on.
Cadence: Channels on a daily two minute scan, full sweep weekly, pay bands re-checked quarterly.
Entry types. Listing is a specific role that will expire. Employer is a careers page, always current. Pool is a filtered search that refreshes itself. The board deliberately leans on the last two so it stays useful between sweeps.
Two links per card where both exist. Open posting goes straight to the specific role. Open site goes to the careers page or search it came from, which stays alive after the posting expires. Cards with a direct link sort to the top and have their own filter chip, since that is the one you actually want.
Why every card does not have one. Job boards expire postings and rewrite URLs within weeks, so a direct link captured today is often dead a fortnight later. Where there is no stable direct URL, the card takes you to the site with the right filter applied and says so, rather than promising a posting and delivering a homepage. Each sweep converts more entries from site-only to direct as real postings appear.