Indigo improves grower profitability, environmental sustainability, and consumer health through the use of natural microbiology and digital technologies. Utilizing beneficial plant microbes and agronomic insights, Indigo works with growers to sustainably produce high quality harvests. The company then connects growers and buyers directly to bring these harvests to market. Working across the supply chain, Indigo is forwarding its mission of harnessing nature to help farmers sustainably feed the planet. The company is headquartered in Boston, MA, with additional offices in Memphis, TN, Research Triangle Park, NC, Sydney, Australia, Buenos Aires, Argentina, and São Paulo, Brazil. http://www.indigoag.com/
As a Data Scientist on the GeoInnovation team focused on leveraging deep learning based algorithms, you will apply your extensive experience in building cutting edge statistical models to complex problems involving incredibly large datasets. You will be both an important individual contributor and a thought leader on the GeoInnovation team. Collaborating across teams, you will also deploy scalable models on our cloud infrastructure and put the results of your work into context for Indigo business leaders.
- Identify role of multi-layered network based models in existing workflows and create development and testing plans
- Take ownership of end-to-end data analyses using these models and deliver on challenging deadlines
- Effectively interface with engineers to scale these models and build code that can be deployed into production systems
- Design long-term strategies to leverage newest developments in machine learning within context of the broader GeoInnovation Data Science priorities
- Contribute thoughtful business recommendations using effective communication of findings (through synthesis and visualization of quantitative information)
- Insatiable hunger to learn; desire to evangelize cutting-edge machine learning techniques, both within Indigo and externally
- Extremely comfortable working with messy datasets from a variety of non-normalized sources
- Enjoys tackling complex analysis problems; able to identify and apply appropriate techniques as necessary
- Proven oral and written communication skills; able to translate results of complex analyses into interpretable results for a variety of audiences
- Strong cross-functional collaboration skills
- Proactive and self-motivated; drives team focus and strategy
- Excellent coding skills, able to inherit and improve an existing codebase; writes well-documented, clean, maintainable code in a collaborative environment
- Thrives in a fast-paced, growth environment
- Excellent prioritization skills
- Basic Qualifications
- 2+ years of professional experience using Python and/or R for data analysis (both preferred).
- Conversant in the use of open-source deep-learning packages to solve real-world problems.
- An academic degree with a quantitative focus.
- Preferred Qualifications
- Extensive practical experience building convolutional, recursive and recurrent neural networks using modern machine learning frameworks such as TensorFlow, Keras, Caffe, Torch/PyTorch or Theano for spatial predictive modelling.
- Extensive practical experience using autoencoders and Boltzmann networks to learn optimal feature embeddings in high-dimensional spaces.
- Experience in mining and managing large volume data-sets.
- Experience with geospatial data, remote sensing and/or imagery data.
- Experience in a software and/or product development environment.
- Experience with git based version control and AWS/Google Cloud infrastructure.