River is constructing "the world's most trusted financial institution to empower people to take ownership of their financial lives through Bitcoin." The organization has secured over $50 million in funding from prominent investors including Goldcrest, Kingsway, Polychain, M13, DG, and Valor. The company demonstrates transparency by publicly sharing financials and proof of reserves.

River seeks a senior engineer comfortable across the full web stack with production system experience at scale. The successful candidate will develop features used daily by clients, working with a tech foundation of Elixir, React, and Postgres.

Hi River team,

The Senior Software Engineer role stands out for pairing React product work with real security and correctness, plus your public proof of reserves and open financials. I am enthusiastic about Bitcoin and your goal of helping people take ownership of their money. I am curious how you split concerns between React and Elixir, how you model custody and transfers, and what guardrails you use for web security in high-risk flows like auth, funding, and withdrawals. I would also love to learn how you validate changes that touch balances and what you publish internally to keep risks visible.

From 2020 to 2025 at Vermont Mutual Insurance, I led the modernization of legacy tools and customer apps into performant React SPAs and drove UI development across teams of 10 to 20 engineers and designers. I partnered closely with design, pushed for UX-first development, and shipped features users praised. I also led a generative AI pilot that evaluated tools, gathered feedback from internal users, and advised on

Call: 20-min preliminary screen with James Page (Chief Product Officer) via Google Meet Role: Senior Software Engineer (React, Full-stack) Not a coding interview — this is a fit/interest conversation

Carving out codectl's role across two systems — as a standalone repo context generator and as contextctl's codebase intelligence engine.

Sources: contextctl research report, repoprompt reverse engineering bible, knowctl repoprompt topic (40 docs), Codex architecture review, competitive landscape research (Aider, Kit, CatCoder, Sourcegraph Cody, Moderne Prethink).

![screenshot](data:image/png;base64,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

Implement the following plan. Return the complete modified files - show the FULL file content for each file, not just diffs or snippets.

Can you explain the algorithm/process we use to sync claude sessions to chromadb? how often? Do we incrementally add conversations? What if a conversations updates? When I semantic search the db how fresh can I expect it to be?

With QMD using ~3GB, you'd have ~5GB free.

For the QMD stack, it'll be great:

QMD models (~3GB VRAM total):