Sequential is building a next-generationAI-drivendiscoveryplatformto identify and designnovelfunctionalactives, including peptides and complex ingredient systems. The platform integrateslarge-scalebiologicaldatasets(>50,000 samples and measurements)spanning multi-omics data, microbiome sequencing, clinical and real-world outcomes. Our goal is to translate biological signals intoactionable compound discovery and optimisation, powering a pipeline across:
- We are looking for a Senior Computational Biologist with ML Engineering background to help build, bridge, and functionalise the link between AI-powered biological discovery and real-world clinical outcomes. This role sits across biological discovery and scalable ML engineering. You will own key parts of the end-to-end architecture from data to model to evaluation to deployment, and work closely with ML engineers and software engineers to productise the platform into client-ready outputs. This is a senior role with significant autonomy and technical ownership.
The Data You Will Work With
The platform is built on a growing dataset of>50,000 biological samples and measurements, including pairedpre- and post-treatment observations. The data includes multiple modalities such as microbiomesequencing (16S rRNA sequencing, ITS sequencing, shotgun metagenomics), Multi-omics (proteomics, lipidomics, metabolomics), Clinical and observational data (treatment exposure, formulation and ingredient combinations, clinical outcomes, patient metadata). Datasets includelongitudinalmeasurements, enabling analysis ofbiologicalresponsetointerventions(e.g., ingredient exposure, treatment, formulation).
Responsibilities
- 1) Build the discovery engine (data > signal > candidate)
- Develop models that identify novel functional actives from multi-omic datasets
- Detect patterns in biological signatures that correlate with clinical outcomes (e.g., inflammation reduction, microbiome restoration, barrier repair, malodour reduction)
- Create robust feature representations from:
- clinical metadata and response data
- SNP and risk features (where relevant)
- 2) Predict mechanism + response
- Build predictive models for:
- molecule–microbe interactions
- clinical response forecasting
- safety and developability scoring
- Translate model outputs into interpretable mechanistic narratives for R&D teams and external partners.
- 3) Design and optimise functional complexes
- Implement multi-objective optimisation and scoring frameworks to balance:
- safety and stability constraints
- manufacturability and cost
- regulatory feasibility
- Support generation of:
- repurposed peptides
- newly discovered natural peptides
- 4) Productionise the AI product launch
- Build end-to-end ML pipeline covering ingestion, training, evaluation and deployment
- Develop APIs/services to serve predictions and ranked candidates into internal tools and client outputs
- Create evaluation harnesses to compare predicted vs. observed validation outcomes
- Implement monitoring and governance: drift, data quality checks, model versioning, auditability
- 5) Collaborate cross-functionally
- Work closely with biology, formulation, and clinical teams to design experiments and validation loops
- Partner with product and commercial teams to shape “client-ready” deliverables (e.g., ranked actives, evidence packs, scientific dossiers)
- Lead and /or partner with ML and software teams to define ownership boundaries code reviews norms, and the path from prototype to a maintained service.
12-Month Mission
Month 0–3: Foundation & Proof of Concept
- Establish harmonised datasets and core data pipelines with dataset versioning, documented schemas, and baseline QC checks
- Deliver feasibility screening models for active discovery with an agreed split strategy (sample, cohort, and time splits) and reporting baseline ranking metrics (e.g., hit-rate@K, NDCG@K)
- Build initial predictive baselines with clear metrics
- Implement multi-criteria scoring + optimisation with a defined objective, weighting strategy, and ablation plan
- Extend predictive models and improve candidate ranking performance
- Develop reproducible experiment tracking and evaluation workflows
Month 6–9: Advanced Validation
- Compare predictions against real validation outputs and refine models
- Improve robustness, interpretability, and governance
- Deliver a performance report suitable for internal and external stakeholders
Month 9–12: Client Readiness & Pilot Launch
- Finalise v1.0 AI product outputs (ranked candidates + evidence summaries)
- Support pilot client projects and accelerate validation turnaround time
- Contribute to the first commercial-ready “hero” functional complex pipeline
What we’re looking for
- Strong Python and experience building ML systems end-to-end with evidence (links to shipped projects, tools, or repos)
- Proven ability to work with large, messy real-world datasets and to define leakage-safe validation splits (e.g., sample, time, cohort)
- Practical knowledge of ML evaluation, validation strategy, and failure modes including error analysis and iteration planning
- Experience with model development using PyTorch / TensorFlow / JAX
- Ability to communicate clearly across technical + scientific stakeholders
- Strong experience leading or coordinating cross-functional teams, and thinking strategically across product and science (prioritisation, tradeoffs, and shipping)
- Comfort deploying models (batch + real-time inference) into production environments or owning the handoff to engineering with clear interfaces
Strongly preferred
- Experience with biological or high-dimensional scientific datasets (omics, imaging, microbiome, clinical)
- Experience with multi-modal learning, embeddings, or representation learning
- Familiarity with optimisation / ranking / multi-objective scoring systems
- Interest in mechanism-driven modelling and scientific interpretability
- Foundation model experience (LLMs and/or biological foundation models) with proof (fine-tuning/adapters or applied FM pipelines, plus how it was evaluated)
- A strong portfolio of tool or pipeline development (internal platforms, evaluation harnesses, reproducible workflows)
Nice to have
- Experience building developer-friendly tooling: CLIs, dashboards, APIs, or reusable libraries used by other scientists/engineers
- Some front-end/product sense (you care about shipping usable tools, not just notebooks)
- Experience with causal inference, Bayesian methods, or mechanistic simulation
#J-18808-Ljbffr