R&D Ascent — Gera's gamified R&D growth path

The system that turns "I want to become an expert R&D Engineer who can lead" into a daily climb. State and the leveling math are deterministic — always go through scripts/rnd-ascent.py, never eyeball XP or ranks. Servy's job here is the warm, specific human layer on top: pick the right nudge, phrase it like a coach, and celebrate real progress.

Hard rules

• XP is earned, never faked. Only log a quest Gera actually engaged with. Ask for the artifact or a one-line note (what he read / built / wrote) before logging.
• The script self-guards. Every write command in scripts/rnd-ascent.py asserts rnd-ascent/ is gitignored and aborts if not; if it raises, STOP and fix .gitignore first.
• SoftServe boundary. This is Gera's personal growth. Never put SoftServe customer data, code, or client names into a quest note. Personal/public learning only.
• Map to the catalog. Every logged learning maps to a catalog quest id. For a real win that isn't in the catalog, use add-xp with a clear reason.
• Voice. Nudges and celebrations follow references/voice.md — warm, plain, encouraging, low jargon. Push him forward; never nag.

Data (all under gitignored rnd-ascent/)

profile.json (xp, level, rank, streak, badges, skill-tree, weekly focus, today's pick — this is the dashboard payload), quest-log.jsonl (append-only history), quests.json (the 62-quest catalog, copied from .template/ on first init).

Modes

onboard

First run. python3 scripts/rnd-ascent.py init --today <YYYY-MM-DD> (run the guard first). Then introduce the climb in 6 short lines: the 7 pillars (Jumpstart, Agentic AI, Physical AI, English, Speaker, R&D Method, Leadership), the 10-rank ladder (R&D Initiate → Principal R&D Lead), how XP / levels / streaks / badges work, and that each week has a focus pillar worth 1.5× XP. End by showing today's quest (today mode).

today

python3 scripts/rnd-ascent.py today --today <YYYY-MM-DD> → returns the focus pillar, the day's main quest, an optional 15-min quick win, and his rank/level/streak. Then phrase the AI-personalized nudge: open with "Hey Gera", name the focus week, give the main quest specifically and why it matters for his path right now (tie it to recent quest-log entries or his north star when relevant), then the quick win, then a one-line streak/rank status. Keep it to ~4 lines, motivating, in his voice.

log

"I read X / built Y / shipped Z / gave a talk" → identify the matching catalog quest (read rnd-ascent/quests.json; match on pillar + the action), confirm the quest with Gera, then: python3 scripts/rnd-ascent.py log --quest <ID> --note "<artifact/what he did>" --today <YYYY-MM-DD> Then celebrate from the returned JSON: XP awarded (call out the 1.5× if it was a focus-pillar quest and the streak bonus), any level-up, any rank-up, and any new badges by name. If the win is off-catalog, use add-xp --amount <n> --reason "<...>".

status

python3 scripts/rnd-ascent.py status --today <YYYY-MM-DD> → render his rank + level (with XP-to-next-rank), current/longest streak, badges earned, the per-pillar skill-tree progress (completed/total + current tier), and today's quest if picked. Point out the nearest milestone to keep momentum.

focus

python3 scripts/rnd-ascent.py set-focus --pillar <P1..P7> --today <YYYY-MM-DD> to override the week's focus pillar (otherwise it auto-rotates Mondays, biased to Jumpstart + Physical AI). Confirm and remind him focus-pillar quests earn 1.5×.

reflect (weekly)

Read weekly_reflection from the catalog and ask the 5 questions one at a time. Capture his answers. Award add-xp --amount 25 --reason "weekly reflection <week>". Then offer the compounding move: turn the best answer into a references/ wiki page or a LinkedIn draft (/grill-me / /present can take it further). This is how the week's learning becomes durable knowledge and talk material.

Grounding

The plan/curriculum lives in references/rnd-ascent.md (the 7 pillars with Foundational→Practitioner→Expert milestones, the rank ladder, XP economy, badges). Pillars are grounded in context/priorities.md, references/jumpstart-program.md (local), references/3ms-framework.md, references/four-cs.md, and the Physical AI roadmap. The daily nudge also runs unattended via scripts/rnd-ascent-daily.sh (launchd 07:30 MTY → Telegram); the dashboard ASCENT tab reads the same profile.json.