Role Overview
We are hiring a Full‑Stack Engineer to build the first working version ofCliply’s product end‑to‑end. This role sits across backend, frontend, datamodelling and AI integration. You will work closely with our AI engineers toturn multimodal outputs (video perception, transcripts, reasoning, embeddings)into a usable product surface.
This is not a narrow backend orfrontend role. It is a founding full‑stack role for someone who has shippedcomplete web products independently and is comfortable owning both thetechnical and product behaviour.
What You Will Build (First 2–3Months)
1. Ingest Path
- Register a licensed film or long‑form video.
- Extract a playback proxy (lower‑resolution, streamable).
- Maintain an exact timestamp mapping between source and proxy.
2. Subtitle Handling
- Parse caption files (SRT, WebVTT).
- Store every word with start/end timestamps in Postgres.
- Maintain raw text separately from searchable text.
3. Provider Adapter
- Send video to a managed video‑search or multimodal API.
- Poll for readiness.
- Convert provider output into Cliply’s internal candidate format.
4. Search Page
- Query box with autocomplete.
- List of search results with timestamps.
- Video player that seeks to the exact moment.
- Subtitle line that triggered the match shown underneath.
5. Correction Loop
- Editor UI to adjust in/out points.
- Accept/reject candidate results with reasons.
- Persist corrections and editorial decisions.
Two Non‑Negotiables
- Deletion propagation: removing an asset or revoking rights must remove it everywhere immediately.
- Search correctness: a search that finds nothing must return nothing — never “closest match”.
Key Responsibilities
Full‑Stack Product Development
- Build backend services using Python + FastAPI.
- Build frontend interfaces using React + TypeScript.
- Create clean, intuitive UI without a designer.
- Own end‑to‑end product behaviour from ingest to correction.
Backend Architecture &Pipelines
- Design REST/gRPC APIs for ingestion, retrieval, metadata access and multimodal outputs.
- Build background workers and long‑running jobs with retries, idempotency and resumability.
- Integrate slow, asynchronous or imperfect third‑party AI APIs gracefully.
- Implement deletion propagation, rights revocation and consistency guarantees.
Data & Storage
- Design Postgres schemas for subtitles, metadata, candidates and editorial corrections.
- Build indexes for fast retrieval across multimodal outputs.
- Store embeddings, timelines and metadata graphs efficiently.
Product Iteration
- Sit with users, observe friction, and change the product rather than explain it.
- Work closely with AI engineers to validate perception, alignment and reasoning outputs.
- Deliver investor‑ready demos and internal prototypes.
Collaboration
- Partner with AI engineers on multimodal pipeline integration.
- Partner with backend engineers on scaling ingestion and orchestration.
- Provide early frontend foundations until a dedicated frontend engineer joins.
Requirements
Must‑Have
- You have built and shipped a complete web product end‑to‑end (backend, database, frontend, deployment).
- Strong experience with Python (FastAPI) and React (TypeScript).
- Deep comfort with Postgres: schema design, indexing, query optimisation.
- Experience with long‑running background jobs and understanding of retries, idempotency and resumability.
- Experience consuming slow, asynchronous or unreliable third‑party APIs.
- Ability to build clean, obvious interfaces without a designer.
- Strong product instincts: you iterate based on user behaviour, not assumptions.
- Comfortable working in a fast‑moving pre‑seed environment.
Nice‑to‑Have
- Experience with video processing (ffmpeg, HLS, proxies).
- Experience with multimodal AI APIs (vision, speech, embeddings).
- Experience with search systems (Elasticsearch, pgvector).
- Experience with media or broadcast workflows.
- Experience with cloud deployment (AWS/Azure/GCP).