Machine Learning & Operations Advisory Consultant

Formation / Diplômes

Machine Learning and Operations Consultancy Organization The Alliance of Diversity International and CAT drives research-driven innovations that leverage agricultural biodiversity to foster sustainable food systems, ultimately enhancing livelihoods worldwide. Alliance solutions tackle pressing global issues such as malnutrition, climate change, biodiversity loss, and environmental degradation. The Alliance cultivates groundbreaking collaborations to produce robust evidence and integrate transformative innovations, thereby reshaping food systems and landscapes to sustain the planet, foster economic growth, and ensure food security amid a climate crisis. The Alliance stands as a member of CGIAR, an international collaborative research network dedicated to advancing food security worldwide. Background The NAZI (NLP to Develop and Innovate Zero-shot Intelligence) project is in search of a committed Machine Learning Operations (Flops) Consultant to facilitate the development, deployment, and operationalization of machine learning systems that drive SIKA. As a leading voice-first and multimodal AI platform, SIKA specializes in conversational data collection and analysis. Additionally, the consultancy provides support for AI-driven research workflows. The consultant will collaborate across speech, natural language processing (NLP), multimodal AI, and agentic pipelines, facilitating the transformation of models and data systems from research prototypes into robust, scalable solutions optimized for real-world field deployment. To facilitate the integration of formal breeding processes conducted in controlled experimental trials with practical on-farm environments, these systems are being engineered for crop improvement initiatives. Principal use cases include:

• Identifying farmer preferences.
• Assessing plant disease occurrence and scoring.
• Supporting environmental response, particularly climate adaptation. Deliverables generated through this consultancy will be applied to fulfill the specified use cases. About the position The consultant will provide guidance and assistance throughout the entire Flops process, from initial data ingestion and validation to dataset versioning, model training, evaluation, deployment, ongoing monitoring, and iterative improvement. The role encompasses responsibilities across both cloud-based and edge computing infrastructures. The position requires seamless coordination with research machine learning, software engineering, product, and field teams to guarantee system reliability, scalability, and alignment with project objectives. The consultant will facilitate the seamless incorporation of machine learning systems into the SIKA platform, with a focus on deployment workflows that establish connections between:
• Mobile applications.
• Cloud infrastructure.
• Speech and multimodal pipelines.
• Disease detection and severity scoring workflows.
• Backend services.
• FAIRGrounds-integrated systems. The position will also play a key role in enhancing established protocols and methodologies to ensure organizational excellence and operational efficiency.
• Experiment tracking.
• Model governance.
• CI/CD workflows.
• Deployment automation.
• ML system monitoring across NDIZI infrastructure. This is a full-time consultancy role spanning 11 months, with the possibility of extension, to be conducted at the Alliance office in Arusha, Tanzania. Key activities and specific terms of reference Speech and NLP systems refinement The consultant will: Facilitate the seamless deployment and ongoing optimization of multilingual Automatic Speech Recognition (ASR) systems across both cloud-based and mobile platforms. Develop and implement streamlined workflows encompassing the ingestion of speech data, transcription processes, rigorous evaluation, and ongoing enhancements to the model. Develop automated retraining and fine-tuning pipelines that utilize newly acquired field data to enhance model performance. Assist in implementing and deploying large language model (LLM)-driven workflows specifically designed for conversational analysis and trait extraction tasks. Monitor model performance, latency, response reliability, and data drift in real-world operational environments to ensure consistent and optimal functionality. Enhance inference workflows to perform efficiently in environments with limited connectivity and constrained resources. Multimodal pipeline development and deployment
• Assist in the creation of multimodal workflows that integrate speech data, transcripts, metadata, and field images gathered via SIKA. Develop and execute workflows to validate, synchronize, store, annotate, and version multimodal datasets. Provide assistance in the development, implementation, and assessment of multimodal and vision-language artificial intelligence models. Design and implement scalable systems to efficiently manage multimodal datasets and model outputs within cloud-based environments. Provide support in benchmarking, ensuring reproducibility, and optimizing multimodal AI pipelines designed for real-world field deployment. Disease detection and severity scoring Support the creation and implementation of artificial intelligence-driven processes for identifying diseases and assessing their severity through the analysis of field images and multimodal datasets. Develop and deploy data and evaluation pipelines to facilitate disease annotation, validation, benchmarking, and ongoing model enhancement. Provide assistance with incorporating disease scoring workflows into the SIKA and ON platforms to facilitate field-based data collection and analysis. MLOps infrastructure and SIKIA integration Design and implement robust CI/CD pipelines specifically tailored for model training, evaluation, and deployment processes to ensure seamless integration, testing, and delivery of machine learning solutions. Responsible for overseeing workflows related to experiment tracking, model registries, and dataset versioning.
• Develop and deploy comprehensive monitoring and logging systems for machine learning services to ensure operational visibility and performance tracking. Collaborate closely with the software development team to embed Retrieval-Augmented Generation (RAG) pipelines within an agentic deployment framework. Support the implementation and deployment of machine learning services across Google Cloud Platform (GCP), Fairgrounds, and associated infrastructure. Facilitate the seamless incorporation of machine learning services into the SIKA platform, ensuring compatibility across mobile applications, backend infrastructure, API frameworks, and cloud-based systems.

Maintain adherence to data governance standards, security protocols, and ethical AI principles. Deliverables and payment schedule Deliverable 1: Inception report and technical workplan Timeline: Month 1, approximately Week 4 The consultant will craft an inception report that delineates the technical methodology, implementation priorities, infrastructure specifications, integration timeline, and a comprehensive 11-month project plan tailored to Flops, speech processing, multimodal, and disease-scoring workflows within the SIKA platform. Honoraria: 12,000,000 Tanzanian Shillings / 589,762 Kenyan Shillings. Model development, assessment, and the establishment of supporting infrastructure constitute the primary focus of this deliverable, encompassing the full lifecycle from initial model creation through rigorous performance evaluation and the deployment of necessary technical frameworks. Month 5, approximately Week 20 The consultant will design and implement foundational Flops infrastructure and workflows, encompassing CI/CD pipelines, experiment tracking mechanisms, dataset versioning solutions, model registries, and monitoring frameworks. The consultant will assist in creating, assessing, and enhancing speech, multimodal, and disease-scoring models within both cloud and edge computing infrastructures. 18,000,000 Tanzanian Shillings / 884,642 Kenyan Shillings. Integrating and deploying the SIKA AI pipeline represents the critical third deliverable in this initiative. This task entails seamlessly incorporating the AI pipeline into existing systems and ensuring its operational readiness for production deployment. Successful completion requires meticulous validation of each component, rigorous testing to confirm performance benchmarks, and a structured rollout strategy to minimize operational disruptions. The integration phase must align with established technical standards, security protocols, and compliance requirements to guarantee a smooth transition from development to live operation. Month 8, approximately Week 32 The consultant will develop and implement comprehensive deployment workflows that seamlessly integrate machine learning services with SIKA mobile applications, APIs, backend systems, and the Fairgrounds-integrated infrastructure. The consultant is responsible for developing operational pipelines for speech processing, conversational analysis, multimodal workflows, and disease-scoring services, ensuring comprehensive deployment documentation and seamless integration support. 22,000,000 Tanzanian Shillings / 1,081,230 Kenyan Shillings. Deliverable 4: Monitoring, evaluation, and optimization framework Month 10, approximately Week 40 The consultant will establish comprehensive monitoring, logging, benchmarking, and evaluation workflows for deployed machine learning systems, encompassing ASR performance tracking, multimodal pipeline assessment, disease-scoring validation, and model drift detection. The consultant will additionally offer strategic recommendations aimed at enhancing optimization, scalability, and the efficiency of field deployment processes. 16,000,000 Tanzanian Shillings / 786,349 Kenyan Shillings. A comprehensive final technical report and handover package will be prepared and delivered, encapsulating all project outcomes, methodologies, and key insights to ensure seamless transition and future reference. Month 11, end of assignment The consultant is tasked with preparing a comprehensive final technical report that outlines executed workflows, deployed infrastructure, system performance metrics, significant insights gained, as well as actionable recommendations for future scalability and upkeep. The consultant is responsible for providing comprehensive final documentation, deployment guides, pipeline configurations, and knowledge-transfer materials to internal teams. Education Candidates must hold a master’s degree in any of the following disciplines:

• Computer science.
• Data science.
• Artificial intelligence.
• Software engineering.

Expérience

• A related field. Experience Candidates should possess a minimum of three years of professional experience in one or more of the following fields:
• Machine learning engineering.
• MLOps.
• Deployment of AI systems. Technical competencies Applicants should have: Proven expertise in developing and overseeing end-to-end machine learning workflows, encompassing model training, deployment, real-time monitoring, and systematic version control.

Strong programming skills in Python.

• Proficiency in machine learning frameworks, including but not limited to PyTorch and TensorFlow, is required. Professionals should possess hands-on experience with leading cloud platforms, including Google Cloud Platform (GCP), Amazon Web Services (AWS), or Microsoft Azure. Proven expertise in natural language processing, speech technologies, conversational AI, or applications

---

**

[Click the Apply button below to apply, and Create my CV to build a CV tailored to this offer, professionally]