Alfie Amayo — Operations Portfolio

About

Mon, 01 Jan 0001 00:00:00 +0000

I am an operations and product manager based in New York.

My father wanted me to become a doctor. I chose startups instead, and have been part of them since I was 21. What drew me in then still drives me now: building things from scratch and solving hard, ambiguous problems without a playbook.

Most of my career has been spent in spirits and beverage. I have always loved the industry because it is grounded in real products, real operations, and real customer experience. Over the years, I have built and run distilleries across Europe and Asia, and the products I helped create have won more than 30 industry awards. But that was never the point. What I love is building something real, then seeing it out in the world.

Discovery Timeline

Mon, 01 Jan 0001 00:00:00 +0000

Cooling System Failure / Discovery Timeline

Before the Run

Two runs preceded Batch 50 and shaped how I interpreted the outcome. Batches 48 and 49 were both Orange Liqueur, each running for 7.5 hours and passed quality control (QC) without issue.

During those runs, I observed that the 2000L still ran hotter than the 300L legacy still, with distillate output temperature progressing from cool to warm to hot. The observation was tactile, not instrumented. I noted it but did not return to review it because both batches passed QC and matched the expected flavour.

Discovery Timeline

Mon, 01 Jan 0001 00:00:00 +0000

Juniper Contamination / Discovery Timeline

Pre-Discovery: Batches Pass Initial QC

Between 11 July and 8 August 2022, four batches were produced using juniper lot PDI-JNP-220620: batch 19 (Gold Gin), batch 20 (Dry Gin), batch 21 (Jason Kong), and batch 22 (Dry Gin). All four passed post-production quality control (QC) and were approved for bottling.

Each batch was compared against a reference bottle pulled from finished goods rather than a fixed baseline. Because the reference changed each time, gradual drift went unnoticed.

Impact and Response

Mon, 01 Jan 0001 00:00:00 +0000

Cooling System Failure / Impact and Response

Quarantined

1 batch

Batch 50 — never shipped

Potential radius

5 batches

Stopped before Batch 51

Customer exposure

Zero

Contained pre-bottling

Response time

Same day

Last reversible point

Affected Batch

One batch was affected. Batch 50, a Seekers Mekong Dry Gin run produced on the 2000L still on 29 May 2023, failed QC and was quarantined. It never shipped. Post-incident analysis later confirmed a temperature drift of 1–2°C/hr, rising by approximately 24°C across the 12-hour run.

Impact and Response

Mon, 01 Jan 0001 00:00:00 +0000

Juniper Contamination / Impact and Response

Impact

Quarantined

715 bottles

4 batches · 3 products

Recovered

~70%

Highest-volume accounts first

Containment

Cambodia Only

No international shipments

Response time

Same Day

Quarantine on confirmed finding

Once the juniper lot was confirmed as the source on 15 August 2022, the blast radius could be defined. Four batches produced between 11 July and 8 August were affected across three products: Seekers Mekong Gold Gin, Seekers Mekong Dry Gin, and Jason Kong Butterfly Pea Gin. In total, 715 bottles were quarantined.

Programme Overview

Mon, 01 Jan 0001 00:00:00 +0000

Red Lantern Dining / Programme Overview

Lead time

7 → 3 weeks

Phase 1 reduction

Cost reduction

29.80%

GLS raw material

Orders placed

GLS ×10 · LYS ×8

Phase 1 duration

13 months

Mar 2024 – Apr 2025

The Programme

Red Lantern Dining is a 14-site food and beverage franchise. No location had the space or equipment to produce drinks on site, so I built a centralised bottled RTD programme: drinks made externally to a fixed specification, bottled, and distributed to all locations from one warehouse. I owned the programme end to end, from recipe development and factory setup through to logistics, cost optimisation, and delivery.

Root Cause Analysis

Mon, 01 Jan 0001 00:00:00 +0000

Cooling System Failure / Root Cause Analysis

Mechanism of Failure

The 2000L still was connected to a cooling reservoir designed for the 300L legacy still. Under sustained load, heat returned to the reservoir faster than it could be dissipated. Temperature increased at approximately 1–2°C/hr, rising from around 30°C → 54°C over 12 hours.

As coolant temperature rose, the system became less effective at condensing vapour back into liquid. More vapour passed through without being properly captured. Distillate output temperature increased, less intended flavour was retained, and the final profile no longer matched the expected standard.

Root Cause Analysis

Mon, 01 Jan 0001 00:00:00 +0000

Juniper Contamination / Five Whys Analysis

Root Cause

The immediate cause was a contaminated juniper lot, PDI-JNP-220620, received from a Vietnam-based supplier on 20 June 2022. The more important failure was structural: there was no intake control capable of detecting it before it entered production.

The exact cause of the contamination was never confirmed. Sensory evidence pointed to either microbial activity or oxidative degradation during storage or transit, but no lab analysis was performed. The investigation confirmed the failure was upstream. The juniper arrived compromised. No error was identified in distillery handling, production, or quality control execution.

Structural Improvements

Mon, 01 Jan 0001 00:00:00 +0000

Cooling System Failure / Structural Improvements

Four changes followed the incident. Each was designed to close a specific failure mode exposed by the investigation.

Detection

Change 1

Coolant reservoir temperature instrumented and monitored

A 50°C threshold is now used for coolant reservoir temperature across all batches. It was set from the highest recorded temperature on the 300L still, which consistently produces spirit that passes quality control. Above that level, cooling becomes less effective and product quality begins to drop. Any breach now triggers immediate investigation while the batch is still active, including checking reservoir temperature and confirming the chiller is operating as expected.

Decision Constraints

Change 2

Hard stop on weak signals before raw material commitment

Any weak signal, sensory or operational, is now enough to halt production before raw materials enter the still. Once added, the batch is committed and cannot be recovered. The stop is now a fixed rule, not a judgement call. This delays irreversible commitments until the issue is understood, while action is still possible.

Change 3

Explicit escalation trigger for threshold breaches

One person still has authority to halt production, but the trigger for review is now fixed rather than decided in the moment. Any threshold breach requires a documented review before the run continues. When one person both sees the signal and makes the decision, weak signals are easier to dismiss. The documented review removes that reliance.

Execution Controls

Change 4

SOP versioning with explicit update loop

When a batch requires a workaround or exposes ambiguity, the SOP is updated before the next batch and the previous version is archived. During the incident, key assumptions, including the belief that visible condensation was a reliable quality signal, were being held in memory rather than written down. SOP versioning records these changes and reduces the risk of reverting to outdated practice.

Structural Improvements

Mon, 01 Jan 0001 00:00:00 +0000

Juniper Contamination / Structural Improvements

Two structural changes came out of this incident. Each was designed to close a specific detection gap identified during the investigation.

Change 1

Incoming lot validation introduced

Every new botanical lot is now tested at intake through single botanical distillation against the house baseline before it enters production. This moved detection upstream to the point where the ingredient could be rejected before use. Lots that match the baseline are approved. Lots that deviate are rejected.

Change 2

Fixed reference batch introduced for quality control

Each batch is now compared against a fixed, documented reference batch. Before this, batches were compared against a reference bottle selected from available stock, so the comparison point changed over time. This allowed gradual flavour drift to pass unnoticed across multiple batches.

Blast Radius Query

The query below traces lot PDI-JNP-220620 through the database from stock receipt to bottled output.

Supply Chain Dashboard

Mon, 01 Jan 0001 00:00:00 +0000

Red Lantern Dining / Supply Chain Dashboard

This dashboard reconstructs the Phase 1 operating history in a way the Excel tracker could not. OTIF is the primary delivery metric. COGs is shown order by order, with each inflection point tied to a specific decision. The data covers Apex Beverage, Crest Bottling, and seven suppliers. All figures are real or reconstructed from production records. Client and supplier names are fictionalised.

Live artefact

Open the Supply Chain Dashboard

React · Vite · Recharts · Deployed on Vercel

↗

Query Panel

The dashboard includes a query panel built around five operational questions. These queries are designed to surface patterns that are not visible order by order: which interventions worked, where the programme remained fragile, and where reliability was chosen over margin.

Systems Context

Mon, 01 Jan 0001 00:00:00 +0000

Cooling System Failure / Systems Context

Operations Overview

Seekers Spirits is a craft distillery in Phnom Penh, Cambodia. At the time of this incident, production ran on two stills: a 300L legacy still and a 2000L still installed in early 2023. The 300L still had a stable, established operating pattern. The 2000L still handled larger runs, but its cooling system depended on a reservoir originally sized for the smaller equipment. I owned quality control end-to-end, from tasting through to final release decisions.

Systems Context

Mon, 01 Jan 0001 00:00:00 +0000

Juniper Contamination / Systems Context

Operations Overview

At the time of this incident, Seekers Spirits produced three core gin SKUs on a single 300L still: Seekers Mekong Dry Gin, Seekers Mekong Gold Gin, and Jason Kong Butterfly Pea Gin. Quality control (QC) was my responsibility end-to-end, from tasting through to final release decisions.

The production flow had no formal quality gates upstream of production and no automated monitoring between stages. Ingredients were received and visually inspected, batches were produced, and finished product was assessed at post-production QC before release. Any failure introduced at ingredient receipt had an unobstructed path through every stage before it could be caught.