WILDLABS
World Wide Fund for Nature/ World Wildlife Fund (WWF)
See the original posting here. Below is a reposting of the original posting.
June 30 update: The original deadline, June 29, has now passed. We have now changed the application deadline to rolling. Note that applicants who apply after the original deadline will be generally less likely to receive a job offer, and won't necessarily hear back from us (though all who have applied by June 29 will).
Is this you?
- 🚀 Are you interested in building one of the world’s first reliable satellite imagery models for water quality prediction?
- 🛠️ Are you a machine learning or remote sensing engineer who loves applying hard technical skills to wild, unsolved problems?
- 🌏 Do you thrive on field visits and enjoy collaborating directly with on-the-ground teams?
- 🌟 Are you eager to use your technical skills to make a difference in the world?
If you answered yes, and have strong machine learning experience (ideally in remote sensing), this role might be for you.
Criteria Notes—Before you read further. . .
- This role is for a 6-month contract, though it is potentially extendable, dependent upon your success and our future needs.
- We are hiring for this role internationally, though our preferred but optional criteria is that this person would spend at least a few weeks at our field site in India (travel and accommodation covered).
- Application Deadline: June 29, 2025. Apply through this form.
Who We Are
Fish Welfare Initiative (FWI) is a nonprofit dedicated to improving animal welfare in aquaculture (fish farming). We currently partner with 160+ fish farms through our Alliance for Responsible Aquaculture and are building tools, like satellite analytics, to reach thousands more.
Your Mission
As part of our R&D work to develop the next generation of our interventions, FWI has been investing for over a year in developing a model that can assess or forecast the water quality at fish farms remotely.
Specifically, we have explored satellite imagery-enabled remote sensing models and forecasting models using existing data, satellite imagery, and weather APIs. To date, we have only worked with part-time, external partners. So far, none of the models have proven accurate enough—this seems to be a technically very challenging project, and more training data may be required than these partners have used previously.
If hired, your mission would be:
- Build and validate robust satellite-based and forecast-driven models that accurately predict real-world water quality parameters: dissolved oxygen, ammonia, pH, and chlorophyll-a.
- Iterate with ground‑truth teams: design data collection protocols; supervise associates gathering in‑pond measurements.
- Share learnings: document code, publish internal memos, and present interim results to leadership.
Resources: Note that you would have access to our team in India, which is adept at collecting ground truth data. For reference, see our publicly shared historical data.
Deliverables
To be useful for FWI’s programs, each model—whether based on remote sensing or forecasting—must meet minimum accuracy thresholds when validated against ground-truth data collected by our team. The specific deliverables and their associated performance benchmarks are listed below. If a deliverable isn’t met, the contract may end early.
Early submission of deliverables is, of course, encouraged. You may receive a bonus of up to $25,000 upon early submission of deliverables (further details to be determined in the contract). In case of early submission, the current contract would need to be reassessed.
Deliverable 1 (end of month 4)
- Preliminary models for any three of the four target parameters (dissolved oxygen, pH, ammonia, chlorophyll-a) achieving a coefficient of determination (R-squared) of ≥ 0.5 when validated against our ground-truthed data set.
Deliverable 2 (end of month 6)
- Models for at least three water quality parameters capable of determining/predicting values with the following minimum criteria when validated against our ground-truthed data set:
Minimum criteria for each model
Statistical Parameter
Coefficient of determination (R-squared)
Root mean square error (RMSE)
Mean absolute error (MAE)
Dissolved Oxygen
≥ 0.66
≤ 1.62 mg/L
≤ 1.07 mg/L
pH
≥ 0.71
≤ 0.13
≤ 0.11
Ammonia
≥ 0.66
≤ 0.08 mg/L
≤ 0.06 mg/L
Chlorophyll-a
≥ 0.73
≤ 59 µg/L
≤ 42 µg/L
For each model, the code, docs, Docker/GEE scripts to be shared with FWI.
Qualifications
Required Qualifications
- Demonstrated ability to build ML models based on complex datasets.
- Fluency in Python and a major ML framework (e.g., TensorFlow, PyTorch, XGBoost).
- Strong statistical experience; comfortable owning an end‑to‑end data pipeline.
- Demonstrated ability to work independently and concisely interpret and communicate code-based work output to non-technical team members.
Ideal Qualifications
- Experience building ML models on satellite and/or remote sensing data.
- Experience with forecasting and predictive models.
- Willingness and ability to spend at least four weeks at our field site in Andhra Pradesh, India.
- Experience with aquaculture, limnology, or water‑quality monitoring.
- Familiarity with Google Earth Engine or open EO APIs.
- Prior work in resource‑constrained or field‑based R&D.
- Interest and/or prior experience with the effective altruism and/or farmed animal welfare movement.
The Fine Print
- Location: Remote with optional but preferred multi-week field visit. Candidates who can work with ≥3 overlapping hours with India Standard Time working hours are preferred.
- Position Status: Full‑time, 6‑month contract. Managed by Paul Monaghan (R&D Lead).
- Compensation: Net pay expected USD 40k-80k for 6 months, adjusted for cost of living and experience; travel stipend also included.
- Start Date: ASAP.
- Benefits: Flexible hours, autonomy, a chance to develop technology where little has existed before.
Equal Opportunity
FWI is committed to building a diverse and inclusive team. We encourage people of every color, orientation, age, sex, gender, ethnicity and ability to apply. Please contact us if you have any questions or concerns regarding accessibility, and we will endeavor to make adjustments throughout the process where possible.
How to Apply
Please complete this form by June 29, 2025.