About Sqwish Labs
At Sqwish Labs we build the infrastructure that lets AI products learn from real production outcomes, so signals like customer resolution, conversion, trust, time saved, safety and cost can shape what the system does next.
The work is hard in the best way. Part of it is research: learning from messy, delayed, real‑world feedback instead of clean benchmarks. Part of it is engineering: building reliable infrastructure that can sit inside live AI systems and make good decisions request by request.
Position Overview
This is a flexible internship in Cambridge that can be remote or in‑person, shaped around one of two tracks: AI Research / Engineering or Backend Engineering. You might work on experiments, evaluations, reward signals, and model behaviour; or on APIs, data pipelines, infrastructure, observability, and production systems that power the learning loop.
Responsibilities
- Prototype and test learning algorithms, including models that help AI systems learn continuously over time from sparse and delayed signals.
- Design reward functions, evaluation metrics, and experiment protocols.
- Analyse model behaviour, failure cases, noisy feedback, and misleading metrics.
- Run experiments with clear hypotheses, reproducible configurations, and honest analysis.
- Design robust APIs and build production‑quality Python and Rust services.
- Work with Postgres, Redis, background jobs, queues/streams, and migrations.
- Add tests, structured logs, traces, metrics, and dashboards for real systems.
- Improve local development, Docker, CI, E2E tests, and release workflows.
- Work with research to move research ideas toward reliable production systems.
Nice to Have (Teachable on the Job)
- Experience with ML, LLMs, embeddings, prompt optimisation, model routing, or evaluation.
- Familiarity with reinforcement learning, contextual bandits, reward modelling, or uncertainty.
- Experience with Python, FastAPI, Pydantic, SQLAlchemy, pytest, or data analysis tools.
- Interest in Rust, low‑latency systems, APIs, databases, or distributed systems.
- Exposure to Docker, Kubernetes, GitHub Actions, observability, or cloud infrastructure.
- Contributions to research projects, open‑source, internal tools, or strong personal projects.
Compensation & Benefits
$2k–$4k/month based on location and experience.
Duration, Timing & Location
Minimum 10–12 weeks (can be longer). Available year‑round (summer, winter, or longer by mutual agreement). Remote or in‑person in Cambridge.
#J-18808-Ljbffr…