Tech Transfer and Scale-Up: Models That Travel Between Scales

📍 Where we are: Part II · Discovery & Development, Learned — Chapter 9. The analytical methods chapter built the chemometric and deep-spectroscopy models that read a spectrum or a chromatogram. This chapter asks the uncomfortable follow-up: do those models still work when you move them to a different probe, a different reactor, a different scale?

Every model in this book so far has been trained and validated on data from one place — one probe, one Ambr mini-bioreactor bank, one development suite, one set of raw-material lots. That is exactly where models are cheapest to build and where the data is densest. It is almost never where the medicine is actually made. The whole point of process development is to hand a recipe to manufacturing, and a model that rode along with that recipe has to make the same journey: from 15 mL to 2,000 L, from a development Raman head to the plant's installed analyzer, from the team that built it to the team that has to trust it on a commercial batch. This chapter is about what happens to a learned model on that journey, and it is mostly a chapter about failure modes — because the default outcome of moving a model is that it gets quietly, sometimes catastrophically, worse.

The honest framing is this: transferability is the binding constraint on machine learning in biomanufacturing, more than algorithm choice or even data volume. A bioprocess model is not a law of nature; it is a fit to a particular measurement apparatus observing a particular living system under particular conditions. Change the apparatus or the conditions and the fit no longer holds — and the model, being just arithmetic, has no way to tell you it has stopped being right.

The simple version

Imagine you learn to judge a soup's saltiness by tasting it in your own kitchen, with your own spoon, on your own tongue. You get very good at it. Then someone hands you a different spoon, in a restaurant kitchen, with the soup made at fifty times the volume in a vat instead of a pot — and asks you to call the salt level to the gram. Your skill does not simply transfer. The spoon tastes different, the big vat is not stirred the same way, and your calibrated tongue was tuned to your soup. A bioprocess model is in exactly this position when it moves from the lab to the plant. This chapter is about the two ways to cope: re-tune your tongue to the new spoon (calibration transfer), or build a model that understands why soup gets salty (mechanistic hybrid) so it travels better in the first place.

What this chapter covers

• The transferability problem framed precisely — why a model trained at small scale degrades at large scale, decomposed into the three shifts that cause it (instrument, scale, and biology).
• Calibration transfer / domain shift for spectra — the concrete machinery (Direct and Piecewise Direct Standardization, transfer standards, Kennard-Stone selection) that re-aligns a spectral model across probes and scales, with real published numbers.
• Hybrid mechanistic scale-down models — using first-principles physics so the part that must be learned is small enough to survive the jump.
• Predicting scale-up risk — turning scale-up from a guess into a model whose inputs are engineering parameters (kLa, mixing time, shear, CO₂ stripping).
• Run-to-run variability — the living-system noise floor that limits all of the above, and why it makes transfer validation mandatory rather than optional.
• A runnable calibration-transfer demonstration over the real Raman dataset, contributed as examples/platform/ml/transfer.py.

The transferability problem, decomposed

When people say "the model degraded at scale," they are usually compressing three distinct failures into one phrase. Pulling them apart is the whole game, because each has a different fix.

Shift 1 — the instrument (the measurement moved). A Raman or NIR soft sensor does not see glucose; it sees photons that a particular probe, on a particular spectrometer, with particular collection optics and laser power, turned into intensities. Two probes of the same model, looking at the same broth, do not produce identical spectra: laser power drifts, the focus differs, the optical window fouls at a different rate, and the wavenumber axis can be off by a fraction of a channel. The chemistry is identical; the number the model reads is not. This is the most measurable shift, and the most fixable.

Shift 2 — the scale (the physics moved). A 15 mL Ambr vessel and a 2,000 L stirred tank are not the same reactor scaled by a constant. Mixing time grows, oxygen transfer (kLa) changes, dissolved CO₂ accumulates because the big tank strips it less efficiently, hydrostatic pressure shifts gas solubility, and shear gradients near the impeller appear that simply do not exist in a mini-bioreactor [1]. The cells genuinely behave differently because their environment is different. A model that learned "this glucose trajectory implies this titer" at small scale is now watching cells living a different life.

Shift 3 — the biology (the system moved). Even at one scale, on one probe, the next batch is not the previous batch. A new working cell bank vial, a new media lot, a seasonal shift in raw materials, an aged inoculum — all move the process. This is run-to-run variability, and unlike the first two it is not a one-time offset you can calibrate away; it is a standing noise floor that every transfer must clear. We return to it at the end because it caps everything before it.

Formally, all three are covariate shift and concept drift: the distribution of inputs P(X) changes (a probe or scale moves the spectra), and sometimes the input→output relationship P(y|X) changes too (the cells respond differently). Classical supervised learning assumes both are fixed between training and deployment. Bioprocess scale-up violates that assumption on purpose, by design, every single time.

Calibration transfer: re-tuning the model to the new instrument

The most tractable shift — and the one with the most mature toolkit — is the instrument shift. The field that owns it is calibration transfer (sometimes model transfer or spectral standardization), borrowed wholesale from analytical chemistry, where the same problem is decades old: a NIR calibration built on one spectrometer must run on another without re-collecting thousands of reference samples.

The cleanest evidence that this is a real, large effect comes from a controlled bioprocess study. Pétillot and colleagues immersed two probes of the same Raman analyzer into the same CHO culture at the same time — deliberately removing all biological and scale variability, leaving only the instrument difference. The cross-probe prediction error on cell density was roughly 20%, purely from probe-to-probe optics. After a calibration-transfer step — Kennard-Stone sampling plus Piecewise Direct Standardization — that error roughly halved, to about 10% [2] (peer-reviewed-independent, research). Read that twice: identical chemistry, identical timing, identical analyzer model, and a frozen calibration was 20% wrong from the spoon alone.

How calibration transfer actually works

The core idea is a function that maps spectra from the new instrument back onto the basis the old model expects, learned from a small set of transfer standards — samples measured on both instruments.

• Direct Standardization (DS) fits a single linear transform F such that X_source ≈ X_target · F over the standards, then applies F to every new spectrum before the frozen model sees it. Powerful, but it lets every wavenumber influence every other — easy to overfit with few standards.
• Piecewise Direct Standardization (PDS) is the workhorse. Instead of one global transform, it fits a local map for each wavenumber j: predict the source intensity at j from a small window of target intensities around j. This respects the physics — a probe distortion at one Raman shift mostly affects nearby shifts — and needs far fewer standards [2].
• Transfer-standard selection decides which samples to measure twice. Kennard-Stone picks a maximally spread-out subset so the standards span the spectral space rather than clustering, which is why it pairs naturally with PDS.

The discipline this imposes is the heart of the chapter: a model that moves to a new probe is, for regulatory purposes, a new analytical procedure until it is re-qualified. ICH Q14 treats a multivariate calibration as a procedure with a lifecycle — defined operating range, documented validation, ongoing performance monitoring — and calibration transfer is documented work with its own evidence, not a free lunch [3] (regulatory). The open-source analytics chapter draws the same line from the soft-sensor side; here we make the transfer step itself runnable.

Calibration transfer in code, on the real Raman dataset

examples/platform/ml/transfer.py makes the whole arc concrete on the real simulated spectra in datasets/raman_spectra.parquet — the 336 hourly Raman spectra of BATCH-2026-001, 701 wavenumbers (wn_400 … wn_1800), paired with the kinetic titer_g_L reference. We train a PLS soft sensor at "small scale," then synthesize a different-probe / larger-scale spectral response — a wavenumber-dependent gain, a sloped additive baseline, and a one-channel axis drift, the canonical instrument distortions — apply the frozen model naively, watch it break, and then recover it with PDS using 24 transfer standards. The synthetic shift is clearly labelled illustrative (we have one real batch on one simulated probe); the pipeline is real scikit-learn and NumPy.

# examples/platform/ml/transfer.py — calibration transfer across a (synthetic) probe shift.
def fit_pds(Xs_std, Xt_std, win=11):
    """Piecewise Direct Standardization from paired transfer standards.
    For each wavenumber j, a local least-squares map predicts the SOURCE
    intensity at j from a window of TARGET intensities around j."""
    n_wn = Xs_std.shape[1]; half = win // 2; maps = []
    for j in range(n_wn):
        lo, hi = max(0, j - half), min(n_wn, j + half + 1)
        A = np.hstack([Xt_std[:, lo:hi], np.ones((Xt_std.shape[0], 1))])  # window + intercept
        coef, *_ = np.linalg.lstsq(A, Xs_std[:, j], rcond=None)
        maps.append((lo, hi, coef))
    return maps

# 1. fit PLS on the source ("small-scale") spectra; 2. score in-domain;
# 3. apply naively to the shifted ("large-scale") spectra; 4. correct with PDS on
#    24 Kennard-Stone-style transfer standards; 5. re-score.
idx  = np.linspace(0, len(X) - 1, 24).astype(int)          # transfer standards
maps = fit_pds(X[idx], X_tgt[idx])                         # learn local maps on standards...
X_corr = apply_pds(maps, X_tgt)                            # ...then correct ALL target rows

Running python platform/ml/transfer.py prints exactly this (verbatim run output):

Calibration transfer demo (illustrative shift) over 336 spectra x 701 wavenumbers, 24 transfer standards
  source  (in-domain) : R2=0.9995  RMSE=0.0356 g/L
  naive transfer      : R2=-577.7308  RMSE=39.1968 g/L   <- model breaks on the new probe
  PDS-corrected       : R2=0.9908  RMSE=0.1567 g/L   <- recovered after calibration transfer
ASSERT ok: PDS calibration transfer reduces cross-probe RMSE.

The story is stark and it is the whole point. In-domain, the model is near-perfect (R² = 0.9995). Move it to the new probe with no transfer step and it does not merely degrade — it produces physically absurd predictions (RMSE of 39 g/L against a titer that never exceeds 5.7 g/L; a negative R² means the model is now worse than just guessing the mean). Apply PDS with two dozen transfer standards and it snaps back to R² = 0.9908, RMSE = 0.157 g/L — usable again. The naive number is exaggerated by the size of our synthetic shift, but its direction is exactly Pétillot's finding: a frozen spectral model on a new instrument is untrustworthy until it is transferred, and a small, well-chosen set of paired standards can rescue it.

The transferability problem and its two cures: a model trained at small scale meets three stacked shifts when it crosses to large scale; calibration transfer re-aligns the measurement using a few paired standards, while a hybrid mechanistic backbone shrinks what has to be learned so the model travels in the first place. Original diagram by the authors, created with AI assistance.

Transfer across scales, not just probes

Calibration transfer was invented for instrument-to-instrument, but the same machinery is what the industry now uses for scale transfer. Bayer reported moving Raman PAT models from Ambr 250 mini-bioreactors up to 10 L stirred tanks, retaining usable predictive performance (glucose R² ≈ 0.84 at the larger scale) by transferring the calibration rather than rebuilding it from scratch [4] (vendor/trade-press-self-reported, pilot). The economic logic is overwhelming: building a calibration at 2,000 L means sacrificing commercial-scale batches to generate paired reference data, at six figures per run. Transferring a model built cheaply at small scale, validated with a handful of large-scale standards, is the only affordable path.

Scale also introduces measurement problems that have no small-scale analogue. Merck's perfusion-Raman platform had to solve fluorescence interference that only appears at the very high cell densities of intensified culture — a spectral artifact the development model never saw because development never ran that dense — before the transferred model could be trusted [5] (peer-reviewed, pilot). That is the difference between covariate shift you can standardize and a new physical phenomenon you must first handle: calibration transfer fixes the former; the latter needs a modeling change.

Hybrid scale-down models: travelling light by knowing physics

Calibration transfer fixes the instrument shift. It does almost nothing for the scale and biology shifts, because those are not measurement artifacts — the cells really are behaving differently. The deeper answer to "make a model that travels" is to make less of the model learnable in the first place.

This is the hybrid (gray-box) idea from the data-management ML chapter and the analytical-methods chapter, now aimed squarely at transferability. A scale-down model is a small-scale model intended to predict large-scale behavior; a hybrid scale-down model keeps a mechanistic backbone — mass balances, Monod-type growth kinetics, gas-transfer relations — that is scale-aware by construction, and uses a learned component only for the residual the physics cannot capture. The kinetic constants are the same biology at any scale; what changes is the environment the balances are solved in, and engineering correlations describe how kLa, mixing time, and CO₂ stripping change with vessel geometry. Feed those scale-dependent terms into the mechanistic part and the model already "knows" that the big tank mixes slower and strips CO₂ worse — knowledge the data never had to supply.

The data-efficiency argument from earlier chapters becomes a transfer argument here. Because the mechanistic part contributes scale-aware structure for free, the learned residual is small, so it can be fit on the handful of batches available at the new scale — and, crucially, it generalizes more safely outside the conditions it saw, because the physics holds the extrapolation in check [6]. A pure black box trained at 15 mL has learned a correlation that is only true at 15 mL; a hybrid has learned biology plus a small correction, and the biology travels. The mechanistic fed-batch simulator behind this book's datasets (examples/sim/bioproc_sim/fed_batch.py) is exactly this kind of backbone: Monod growth, death, lactate inhibition, bolus feeds, and a day-7 temperature excursion, with the kinetic constants exposed so a learned residual can ride on top.

Sartorius has demonstrated this pattern at the pre-commercial scale — a hybrid framework combining first-principles dynamic flux balance, a learned VCD component, and parallel mini-bioreactor data to build predictive digital twins of CHO culture, explicitly aimed at the transfer/scale-up problem [7] (vendor/self-reported, pilot). DataHow, an independent ETH Zurich spin-off (not a Sartorius subsidiary), markets hybrid models with transfer learning as a way to cut the number of experiments a process needs — vendor figures in the 30–60% range, which should be read as vendor/self-reported rather than established fact [8]. The recurring theme is the same: the hybrids that travel are the ones where the data-driven part is kept small and the physics carries the scale.

Predicting scale-up risk before you scale

The most ambitious use of learning here is not to transfer a soft sensor at all — it is to predict, before committing to a large-scale run, whether the process will scale successfully. The features are engineering scale-up parameters: power per volume, impeller tip speed, kLa, mixing time, CO₂ stripping rate, shear exposure. The label is the scale-up outcome — did growth, titer, and the critical quality attributes hold at scale, or did they shift?

A 2024 perspective review in Engineering in Life Sciences surveyed exactly this: ML applied to bioreactor scale-up, treating the jump as a prediction problem informed by computational fluid dynamics and engineering correlations rather than a wet-lab gamble [1] (peer-reviewed-independent, research). The honest status is research, not production: the datasets are small (every large-scale run is precious), the outcomes are multivariate (a process can scale fine for titer but fail for glycosylation), and the same transferability problem recurses one level up — a scale-up risk model built on one molecule's history may not transfer to the next molecule. But the framing is valuable even before the models are deployable, because it forces the question "what physically changes at scale?" to be answered in features a model can use, which is the same question hybrid modeling answers in equations.

For our running example, this is the question that sits between the process-development chapter's Bayesian-optimized design space and the production bioreactor: the conditions that produced the golden BATCH-2026-001 (monomer purity 98.611% by SEC) at development scale carry a predicted risk when transferred to the manufacturing tank, and the OOS sibling BATCH-2026-004 — which failed on host-cell protein (HCP) at 128 ng/mg against a 100 ng/mg limit, not on any spectrally-monitored attribute — is a reminder that the failure can surface in a downstream attribute the upstream model never watched.

Anatomy of a transferred model record

A model that crosses a scale boundary should carry a transfer record, and like every artifact in this series its value is in what travels alongside the numbers. A bare "R² = 0.99" is exactly the claim that calibration transfer exists to puncture: it is true for a probe, a scale, a dataset hash and silently false the moment any of those move. The transferred-model record makes the provenance of the jump auditable.

One transfer is a whole record: the source domain it was trained in (pinned by dataset hash and in-domain metric), the target domain it must work in, the naive-transfer metric that proves it cannot be moved untouched, the calibration-transfer method and standards that rescued it, and the governance fields — requalification, operating range, intended use, lineage — that make the jump defensible. Original diagram by the authors, created with AI assistance.

Read the card top to bottom and the chapter's argument is laid out as fields. The source block is what the analytical-methods chapter hands over: the development probe, the small scale, the training dataset pinned by its MANIFEST.sha256 hash, and the in-domain R² = 0.9995 that looks like success and is in fact only a local truth. The target block is the reality check — the plant probe, 2,000 L, and the naive-transfer metrics (R² = −577.73) that prove the model cannot be moved as-is. The green calibration-transfer block is the work that earns the jump: PDS, 24 Kennard-Stone-selected transfer standards, and the recovered R² = 0.9908. The violet governance panel holds what regulators actually ask for: the requalification status (a transferred model is a new analytical procedure under ICH Q14), the operating range it was re-validated over, the intended-use scope, and the lineage linking the source model version through the transfer event to the target model version. The gap between the first metric and the last is the entire subject of this chapter, drawn as one record.

The unsolved part: run-to-run variability defeats clean transfer

It would be dishonest to present calibration transfer and hybrid modeling as a solved pipeline, because the third shift — biology — does not hold still long enough to be transferred to. Calibration transfer assumes there is a stable target domain to map onto: a probe with consistent optics, a process at a fixed operating point. Bioprocesses violate that. The next batch has a different working-cell-bank vial, a different media lot, a slightly different inoculum age; raw materials drift seasonally; the same 2,000 L tank produces a different "normal" this campaign than last. This is run-to-run variability, and it is not an offset you standardize away once — it is a moving target that every transferred model must keep clearing.

The consequence is brutal for the clean story. The transfer standards you measure today to align the plant probe describe the process as it was today; if the next media lot shifts the spectral background, your standards are already stale. The R² = 0.9908 the demo recovers is conditional on the target domain being the domain you transferred to — and in a living system that domain keeps wandering. This is also why model decay is fast in bioprocess relative to other ML domains: a soft sensor can be in spec at qualification and out of spec two campaigns later with no code change, purely because the biology moved underneath it. The sparse-reference regime from the soft-sensor lifecycle makes it worse: the offline assay that could detect the decay returns only once or twice a day, so the model can drift for hours looking perfectly healthy.

This is the deep reason the field's strongest verdict is hybrid modeling plus transfer/Bayesian priors, not pure ML. Run-to-run variability is precisely the regime where physics-anchored models and informative priors earn their keep: they constrain the model to plausible behavior between sparse reference points, and they degrade gracefully rather than catastrophically when the biology shifts. The 2026 review of transfer learning in bioprocess engineering reaches the same honest place — transfer learning is the near-term workaround for the small-data, high-variability reality, and a true bioprocess "foundation model" that would make transfer trivial does not yet exist as a usable system [9] (peer-reviewed-independent, research). Until it does, every model that crosses a scale boundary must be distrusted on a schedule: re-qualified at the new scale, monitored for drift, and re-transferred when the biology moves.

What this chapter adds to the model suite

This chapter contributes one module to examples/platform/ml/:

• transfer.py — a runnable calibration-transfer / domain-shift demonstration. It trains a PLS titer soft sensor on the real raman_spectra.parquet, synthesizes a different-probe / larger-scale spectral response (multiplicative gain, baseline tilt, axis drift — illustrative, since only one real batch on one probe is available), shows the naive frozen model collapsing on the shifted spectra, and recovers it with Piecewise Direct Standardization fit from 24 Kennard-Stone-style transfer standards. It exposes fit_pds / apply_pds as reusable functions and ends with a CI-style assertion that the transfer step reduces cross-probe RMSE, so the chapter's claim cannot silently rot.

It builds directly on examples/analytics/soft_sensor.py (the in-domain PLS Raman→titer model) — transfer.py is that model meeting a domain shift — and feeds forward to the upstream soft-sensing chapters, where the transferred model is what actually runs on the production bioreactor.

Why it matters

Tech transfer is where a process either becomes a product or stalls, and increasingly the models built during development are part of what must transfer. If a soft sensor that took six months to calibrate at small scale cannot move to the plant, the development investment is stranded — the manufacturing team falls back to the slow offline assay and the real-time control the model promised never arrives. Calibration transfer turns that from a rebuild into a re-alignment with a handful of standards; hybrid scale-down modeling builds models that travel because they understand the physics that scale changes; and scale-up risk prediction starts to turn the most expensive gamble in manufacturing into something a model can inform. The throughline, as always, is honesty about the limit: a model is a fit to a place and a moment, and moving it is real work with real evidence, not a copy operation. Get the transfer discipline right and development knowledge reaches the plant; skip it and the most sophisticated model in the building is wrong in a way no one notices until a batch is.

In the real world

The production-grade pattern today is transfer the calibration, monitor it forever. Sartorius's SIMCA / SIMCA-online MVDA stack and BioPAT Spectro retrofit Raman into Ambr and Biostat STR vessels precisely so a model built at small scale can be moved up the scale ladder under monitoring [10] (vendor, production for monitoring). Amgen's Juncos, Puerto Rico site runs SIMCA OPLS harvest-titer models in commercial GMP — a first-party/self-reported account of in-process models operating at commercial scale, which is exactly a model that survived the transfer to manufacturing [11] (production, first-party self-reported). The published calibration-transfer studies that quantify the problem — Pétillot's two-probe experiment and Bayer's Ambr-to-10 L transfer — are the strongest evidence that the effect is large and the fix is real [2][4].

The regulatory frame is catching up. FDA's 2023 discussion paper Artificial Intelligence in Drug Manufacturing names model maintenance and re-validation across deployment changes as an open question for AI under cGMP [12] (regulatory), and its 2025 draft guidance on AI to support regulatory decision-making frames model credibility as a function of model influence and decision consequence — a transferred model touching a CQA decision sits high on both axes [13] (regulatory). The ISPE 7th Pharma 4.0 survey's blunt summary applies here as much as anywhere: AI/ML has the most pilots and the fewest scaled implementations, and the production clusters are monitoring and human-in-the-loop, not autonomous cross-scale control [14]. Cross-scale model transfer is real and useful; cross-scale autonomous model-driven control is still mostly a slide.

Key terms

• Transferability — the degree to which a model trained in one domain (probe, scale, process) still performs in another; the binding constraint on bioprocess ML.
• Calibration transfer (model transfer) — re-aligning a spectral model to a new instrument or scale using a small set of paired samples, so the frozen model can be reused without full recalibration.
• Transfer standards — samples measured on both the source and target instruments, used to learn the calibration-transfer map.
• Direct Standardization (DS) / Piecewise Direct Standardization (PDS) — the global vs. local linear transforms that map target spectra onto the source basis; PDS fits one local map per wavenumber and needs fewer standards.
• Kennard-Stone selection — an algorithm that picks a maximally spread-out subset of samples as transfer standards, so they span the spectral space.
• Covariate shift / concept drift — the input distribution P(X) changes (a probe/scale moves the spectra) and/or the input→output relationship P(y|X) changes (the cells respond differently).
• Scale-down model — a small-scale model intended to predict large-scale behavior; hybrid scale-down keeps a scale-aware mechanistic backbone so the learned residual is small enough to travel.
• kLa / mixing time / CO₂ stripping — the engineering parameters that change with vessel scale and drive the scale shift; the features of a scale-up risk model and the scale-dependent terms in a hybrid backbone.
• Run-to-run variability — the standing batch-to-batch noise floor from living-system and raw-material variation; not a one-time offset, so it caps transfer and drives fast model decay.
• Requalification — treating a transferred model as a new analytical procedure under ICH Q14, with re-validation, defined operating range, and ongoing monitoring.

Where this leads

The model has been built, validated, and — when it must move — transferred and re-qualified. Now it goes to work upstream. The next chapter, Seed Train: Soft Sensing the Inoculum and Predicting Contamination Risk, opens Part III by applying these learned models to the first production-scale step: estimating the state of the inoculum from sparse signals and predicting the contamination risk that can end a campaign before the production bioreactor is even filled — the first place a transferred model has to earn its trust on a real run.