Samuel Adole

Selected achievements

• 30% / 40%

  Cut backend API latency by 30% through PostgreSQL query optimisation and Redis caching, and lifted data-API throughput by 40% on a high-availability health-data platform.

• 2 platforms

  Architected two greenfield backend platforms end-to-end — fintech and supply-chain SaaS — acting as sole backend lead on the fintech build.

• 10,000+ daily

  Built ETL pipelines processing 10,000+ AI transcription outputs a day into EMR systems, and real-time dashboards over 20,000+ monthly shipment events.

• 80%+ coverage

  Set the engineering quality bar across a microservices estate through code review, CI/CD gates and sustained test coverage, mentoring engineers on testing and observability.

Stack

Languages

Python, TypeScript, JavaScript (ES6+), SQL, Node.js, Bash

Backend & APIs

Django, Django REST Framework, FastAPI, Flask, Falcon, NestJS, Express — REST, WebSockets, OpenAPI / Swagger

System architecture

Distributed systems, microservices, event-driven architecture, multi-tenant SaaS, API contract design, high availability, fault tolerance

Data engineering

PostgreSQL, Redis, MongoDB, Celery, ETL pipelines, schema design, query optimisation, data reconciliation

Cloud & DevOps

AWS (EC2, S3, Lambda, RDS), Docker, Kubernetes, Helm, Ansible, Linux, GitHub Actions CI/CD, Vercel, infrastructure-as-code

Testing & observability

PyTest, Jest, React Testing Library, test-driven development (TDD), structured logging, CI/CD quality gates, code review

Security & compliance

JWT, role-based access control (RBAC), secure REST API design, audit trails, healthcare data compliance

Frontend

React, Next.js, Redux Toolkit — component-driven UI, real-time data visualisation

AI & LLM engineering

LangChain, Pydantic AI, OpenAI API, Anthropic Claude API — RAG, embeddings and vector retrieval, prompt design, tool / function calling, schema-validated output

LLM eval & observability

RAGAS (retrieval relevance, groundedness, answer correctness), custom pipeline tracing and instrumentation, ChromaDB, BM25, Reciprocal Rank Fusion

Machine learning

PyTorch, TensorFlow, STT / ASR training pipelines, dataset preparation and augmentation, NLP, statistical modelling

Leadership

System-design ownership, engineering standards, mentoring, Agile / Scrum, technical writing, incident response

Experience

1. Jun 2026 — Present

   AI Generalist

   Mercor · US (remote)

   Global AI talent platform for model evaluation and developer placement.

   • Evaluated 500+ pieces of AI-generated content and corporate presentation decks, setting strict quality criteria to filter out repetitive language, hallucinations and unpolished layouts.
   • Analysed complex technical prompts and gave human-in-the-loop feedback to fine-tune frontier LLMs' reasoning and coding.

2. Oct 2023 — May 2026

   Senior Full-Stack Software Engineer

   Code & Care · Remote

   Software development & out-staffing company — embedded in client engineering teams.

   TOCA Social

   Oct 2023 — Nov 2025 · London, UK (remote)

   • Designed and maintained public REST APIs and WebSocket services (Node.js / TypeScript) powering real-time, multi-venue synchronisation across a distributed microservices architecture on AWS — up to 10,000 concurrent requests per second with weekend event peaks, against AWS service SLAs of 99.5–99.999% — using Docker, Kubernetes (EKS) and Helm for high-availability container orchestration.
   • Owned the TOCA Connects integration layer end to end: a cross-studio data-synchronisation service connecting internal platform services to third-party venue-management systems, with the reconciliation and consistency guarantees enterprise integrations demand.
   • Cut backend API latency by 30% through PostgreSQL query optimisation — custom indexes, refactored joins and Common Table Expressions (CTEs) — alongside Redis caching; introduced structured logging and monitoring that measurably shortened mean time to detection on production incidents.
   • Set and upheld engineering standards through code review and CI/CD quality gates (GitHub Actions), sustaining 80%+ test coverage (Jest, React Testing Library) and preventing regressions at scale.

3. Aug 2024 — Sep 2024

   Python LLM Engineer

   Turing · US (remote)

   AI-powered talent platform — embedded in client engineering teams.

   • Assessed and benchmarked LLM-generated code in Python, TypeScript and advanced SQL (query execution plans, recursive CTEs, normalisation) for functional correctness, algorithmic complexity and architectural design.
   • Built complex software-engineering prompts and gold-standard solutions for evaluation datasets used to fine-tune frontier models.

4. Apr 2022 — Aug 2023

   Senior Full-Stack Software Engineer

   Code & Care · Remote

   Software development & out-staffing company — embedded in client engineering teams.

   Fulton Market Group (FMG)

   Apr 2022 — Aug 2023 · NSW, Australia (remote)

   • Architected and built high-throughput internal REST APIs in Python (Falcon) for a multi-tenant supply-chain SaaS platform, with strict validation, error handling and full OpenAPI documentation.
   • Delivered real-time operational dashboards joining WebSocket-driven Python backends to a Next.js and React frontend, using Redux Toolkit for complex state, giving operations teams live visibility of 20,000+ shipment events a month.
   • Introduced structured logging, polling fallbacks and fault-tolerance patterns across distributed services, significantly reducing on-call incidents.

5. Nov 2021 — Mar 2022

   Full-Stack Software Engineer

   Intron Health · California, US (remote)

   Healthcare technology — EMR integration under compliance constraints.

   • Built high-throughput ETL pipelines processing 10,000+ daily AI transcription outputs into downstream EMR systems via secure REST APIs; designed normalised PostgreSQL schemas and SQL stored procedures for ACID compliance, data accuracy and secure audit trails under healthcare regulations.
   • Developed preprocessing and data-augmentation tooling for STT/ASR model training (PyTorch, TensorFlow), capturing 300+ accent profiles with measurable gains in accuracy and pipeline repeatability.
   • Wrote integration specs covering data flow, failure modes and rollback behaviour — a reliable reference for asynchronous review across time zones.

6. Aug 2021 — Dec 2021

   Full-Stack Software Engineer

   Amenta Media AB (Likvio) · Gothenburg, Sweden (remote)

   Fintech SaaS — greenfield Django / FastAPI platform build.

   • Architected Likvio's backend from scratch as sole backend lead — defining the data model, service boundaries and integration architecture on FastAPI and Django REST Framework — driving test-driven development with PyTest for robust error handling and financial data reconciliation, through to a launch supporting $4,000+ in monthly transactions.
   • Integrated the Fortnox financial API to automate contract lifecycle and invoice synchronisation, and managed high-volume background processing (Celery, PostgreSQL) with retry and reconciliation logic, SQL transaction management, row-level locking and optimised bulk inserts — holding 100% data integrity across concurrent financial operations.
   • Authored complete Swagger / OpenAPI documentation covering endpoints, authentication flows and error schemas, so external developers could integrate on their own.
   • Onboarded and mentored junior developers, establishing coding standards and code-review practice that brought them to independent contribution.

7. Jan 2021 — Apr 2021

   Software Engineer

   V3C Technologies · Abuja, Nigeria

   Multi-role education platform.

   • Designed Flask REST APIs with JWT authentication and role-based access control across teacher, parent and student portals — the same access model enterprise SaaS user provisioning runs on.
   • Standardised API lifecycle documentation (Swagger/OpenAPI) and made architectural improvements that cut developer integration time by 30%.

8. Dec 2019 — Jan 2021

   Full-Stack Python Developer

   eHealth4everyone · Abuja, Nigeria

   Health-data / DHIS2 platform — regulated public-health domain.

   • Built and managed a DHIS2 hosting platform (Django, Ansible), automating server provisioning and configuration management as infrastructure-as-code in a public-health data context.
   • Designed and deployed high-availability data APIs (Django REST Framework, Redis, AWS EC2/RDS) for the HealthThink dataset; optimised SQL query sets, Redis caching and auto-scaling improved overall throughput by 40%.

Selected projects

• 2026

  Real-time observability & evaluation platform for RAG systems

  MSc Data Science dissertation — Leeds Beckett University

  An observability platform in Python 3.12 that instruments Retrieval-Augmented Generation (RAG) pipelines end to end, using LangChain, Pydantic AI and the OpenAI API to capture the retrieval and generation stages as structured traces. The retrieval layer fuses dense and sparse search — ChromaDB (HNSW, cosine, 384-dim all-MiniLM-L6-v2 embeddings) combined with BM25Okapi through a custom Reciprocal Rank Fusion layer (k=60, 0.6 dense / 0.4 sparse).

  Quality is scored automatically with RAGAS 0.4 (context relevance, faithfulness, answer correctness) using Qwen2.5 7B as a fixed third-model-family judge, surfaced as live per-query metrics rather than batch-scored after the fact. After evaluating LangSmith, Arize Phoenix, TruLens and OpenTelemetry's LLM extensions, I built custom instrumentation: a perf_counter timer across four stages — embed, retrieve, generate, evaluate — writing one orjson document per query with token counts reconciled against tiktoken. That decomposition showed RAGAS evaluation alone accounts for 85.7% of end-to-end latency.

  Python 3.12LangChainPydantic AIChromaDBBM25 + RRFRAGASinstrumentation

• 2021 — 2022

  EMR integration pipeline Intron Health

  High-throughput ETL pipelines moving 10,000+ daily AI transcription outputs into clinical EMR systems, built on normalised PostgreSQL schemas and stored procedures for ACID compliance and secure audit trails — plus preprocessing and data-augmentation tooling for STT/ASR training across 300+ accent profiles.

  PythonPostgreSQLPyTorchETL

Education & certification

1. Sep 2025 — Sep 2026

   MSc Data Science

   Leeds Beckett University, UK

   Machine learning, statistical modelling, NLP and algorithmic complexity, applied to real-world data engineering and AI systems.

2. Sep 2014 — Sep 2018

   BSc Computer Science

   Bingham University Karu, Nigeria

3. Jul 2026

   Neo4j Fundamentals

   Neo4j GraphAcademy

   Verify certificate ↗

4. Certification

   AWS Machine Learning Foundations

   Amazon Web Services