Sanctuary Computer: LLM Data Engineer

Headquarters: New York Metropolis

URL: https://sanctuary.computer

Sanctuary Pc is hiring a Distant LLM Knowledge Engineer

Authentic job submit hyperlink:

LLM Data Engineer Original Job Post Link

About garden3d

We’re a worker-owned artistic collective, innovating on the whole lot from manufacturers and IRL communities to IoT gadgets and cross-platform apps. We share revenue, open supply the whole lot, spin out new companies, and spend money on thrilling concepts by means of monetary and/or in-kind contributions.

Our consumer roster contains Google, Stripe, Figma, Hinge, Black Socialists in America, ACLU, Pratt, Parsons, Mozilla, The Nobel Prize, MIT, Gnosis, Etsy & Gagosian. We’re the software program group behind modern merchandise like The Mild Cellphone & Mill, and we a worldwide, decentralized neighborhood area collective referred to as Index House.

We consider our garden3d as collective for artistic individuals, prioritizing a contented, proficient, and various studio tradition. We work on tasks that convey worth to our world, and we steadiness deep take care of the work we do with a real curiosity about life exterior of our jobs.

 Who we’re on the lookout for:

Proper now, we’re on the lookout for a Knowledge Engineer with experience in knowledge pipelines, workflow orchestration, and knowledge integration. Our present method to knowledge ingestion depends on an in-house software that defines light-weight workflows in code and executes them with queues. We’re now increasing this course of by adopting Prefect as an orchestration device, enabling us to handle pipelines for structured and unstructured knowledge from various sources, together with internet crawls and scrapers.

On this position, you’ll work on quite a lot of consumer tasks to search out cost-effective, high-quality, pragmatic options to advanced issues. Obligations will embrace:

• Monitoring and sustaining knowledge pipelines, troubleshooting new errors, and addressing format drift
• Extracting and enriching further knowledge parts from various sources
• Reprocessing and validating giant datasets in batch workflows
• Designing and integrating new knowledge sources into present pipelines
• Aligning and integrating extracted knowledge with the core software knowledge mannequin to make sure consistency and value
• Taking part in code critiques, offering constructive suggestions to teammates and making certain adherence to greatest practices
• Contributing to challenge success by conserving a detailed eye on group velocity, challenge scope, finances, and timeline
• Negotiating with shoppers to align challenge scope with finances and timeline, if wanted

The particular person we’re on the lookout for is blissful, relaxed and straightforward to get together with. They’re versatile on something besides conceits that may decrease their normally excellent work high quality. They work “good”, by fastidiously managing their workflow and staggering options which have dependencies intelligently, they like deep work however are OK coming as much as the floor at times for top-level/strategic conversations.

We imagine individuals with backgrounds or pursuits in design, artwork, music, meals or vogue are likely to have a well-rounded sense of design & high quality — so quite a lot of hobbies or facet tasks is a giant good to have!

Compensation

Our pay scale ranges from $85 p/hr to $130 p/hr pending seniority (& group management expertise), and our tasks are not often lower than 8 full-time weeks at 40 hours per week. Moreover, we pay discretionary bonuses for going over and above — like coaching & teaching others, profitable new enterprise, talking at conferences, and so on.

We desire long-standing relationships with extremely accountable and communicative group members, so we encourage candidates to count on longer-term engagements. A Knowledge Engineer working 40 — 45 full-time weeks might take dwelling $150k — $200k USD.

 Should Have Competencies:

• Robust proficiency in Python
• Expertise with knowledge/workflow orchestration instruments (e.g., Prefect, Dagster, Airflow)
• A radical understanding ETL & knowledge transformation for the ingestion of trade commonplace LLMs (OpenAI, Claude, and so on)
• Familiarity with Massive Language Fashions (LLMs)
• Expert in interfacing with APIs (OpenAI, Google Gemini/Vertex, and so on.) utilizing wrapper libraries similar to Teacher, LiteLLM, and so on.
• Sensible expertise in immediate engineering
• Means to work with structured outputs and doubtlessly device calling
• 5+ years of basic expertise in backend (Ruby on Rails, Elixir Phoenix, Python Django, or Node Categorical) and/or native app improvement (React Native, Flutter, Android, AOSP, Kotlin/Java).

 Good to Have Competencies:

We’re at all times pitching for brand spanking new and thrilling know-how niches. A few of the areas beneath are related to us!

• Expertise with Google Cloud Platform (GCP), notably Cloud Run and Cloud Duties
• Data of search applied sciences, together with embeddings and vector databases for semantic search, in addition to keyword-based search (BM25)
• Familiarity with PySpark for batch knowledge processing
• Expertise working with LLMs, Vector Databases, and different generalist AI-enabled software patterns
• Consumer-facing expertise: working straight with clients to collect necessities and supply technical options
• Product administration expertise: defining product roadmaps and collaborating intently with stakeholders
• Engineering administration expertise: main groups, setting technical path, and mentoring builders

 How we interview:

Our interview course of begins with a name the place you get to satisfy a number of members of our group. From there we’ll ask applicable candidates to participate in a technical train which helps illustrate talent degree and luxury.

→ It’s additionally an effective way to see what it’s wish to work with us and assist help people who might not have the ‘proper title’ however have the expertise and technical know-how for the position.

‍ How we work:

We imagine that there’s a greater steadiness between the poles of freelancing & full time, and for that motive Sanctuary works in another way to most outlets:

• Transparency & Ownership: We launch out Revenue & Loss statements to the neighborhood annually, open supply our greatest concepts, and speak enterprise & cash with everybody within the firm. We’re proud to run our enterprise with integrity, and for that motive we share the whole lot with our group & neighborhood.
• 150% Carbon Negative: Our studio offsets 150% of the carbon we use to do enterprise annually, dated again to our founding in 2015. We flip down work that isn’t consistent with our morals, and we encourage our friends to do the identical. We were certified climate neutral earlier this year.
• Strong Morals: Since our founding, we have turned down someplace between $1mm – $2mm of labor that did not meet our ethical requirements. (Most of that was DTC manufacturers that may’t present a legitimate sustainability initiative).
• Async & Decentralized: We use instruments optimized for calm, considerate communication, and go for async every time potential. We struggle onerous to keep up our focus time.
• Distant Pleasant: Our firm is fluent in distant work, making our office extra decentralized, and democratized within the course of.
• Concepts & Merchandise: In our spare studio time, we work to construct our personal open supply or inside merchandise to diversify & bolster our earnings. We create wonderful know-how merchandise for our shoppers, so why not for the studio?

→ Learn extra on our Substack, over here, or our Medium, over there.

Fast tip! Including a Loom recording to your profile in our kind to showcase your skillset can actually make your software stand out!

Please click on this Necessary hyperlink beneath or copy the URL (Similar because the apply button) to proceed along with your software by telling us a bit extra about your curiosity within the position:

garden3d Creative Form

To use: https://weworkremotely.com/remote-jobs/sanctuary-computer-llm-data-engineer

Source link