Kelyn Njeri
Résumé

Kelyn Paul Njeri

Research Engineer · CTO & Co-founder, DataQRL

Nairobi, Kenya · kelyn.njeri@gmail.com

Kelyn Paul Njeri
Download PDF

Backend and platform engineer with 9 years of engineering experience, and the author of the research software that work now runs under. Distributed systems, payments and data platforms across fintech, logistics and compliance; causal inference and computational biology as published packages. Technical co-founder and CTO at DataQRL, part-time. Almost all of it remote.

Experience

DataQRL
Technical Co-founder & Chief Technology Officer (Part-time)
Mar 2025 – Present
Hamburg, Germany (Remote)
  • Designed and built the platform across eight modules: ingestion, transaction reconciliation, identity resolution, monitoring and alerting, case management, and a dedicated privacy plane.
  • Built entity resolution at scale with Splink and hierarchical fidelity scoring, so parties link across fragmented source systems and the networks single-record screening misses become visible.
  • Deployed anomaly-detection models and a rule engine over Celery and Redis, with connectors for Postgres, MySQL, MongoDB, Snowflake and BigQuery.
  • Built multi-agent orchestration on a Neo4j knowledge graph with LangGraph and Qdrant semantic recall, wrote the high-compute engine in Go, and mapped compliance controls to GDPR, HIPAA and SOX.
  • 2,335 commits since March 2025. Leads a three-person team and recruited the frontend engineer.
Backbone Technologies
Founding Engineer (Contract)
Oct 2022 – Present
Memphis, USA (Remote)
  • One of two engineers on an SME fintech platform, owning the backend, the AWS architecture and the payments stack, working directly with the CEO.
  • Built the event-driven backend in TypeScript and Node.js on BullMQ with Express, MongoDB and Redis, holding p95 API latency under 100 ms.
  • Led the payments architecture: selected Stripe when Chase discontinued WePay, ran the migration, and rebuilt payment processing and payout flows around it.
  • Integrated Plaid for bank-account linking and transaction ingestion, and Modern Treasury for inter-bank payment rails.
  • Architected the platform on AWS and cut monthly cloud spend by 30% in a targeted cost-optimisation pass.
  • Shipped a user-facing ML classifier categorising Plaid transactions for accounting, plus the staff and customer dashboards, instrumented with PostHog.
PowWater
Senior Software Developer
Mar 2024 – Feb 2025
Nairobi, Kenya (Remote-first)
  • Automated dispatch across fill-site assignment, route optimisation and fleet management, cutting delivery time from 2–3 hours to roughly 45 minutes and raising driver utilisation.
  • Introduced data-driven load balancing across high-volume dispatchers, prioritising internal drivers while allocating third-party capacity fairly.
  • Redesigned the logistics platform front end in React with real-time Google Maps tracking.
  • Added unit, integration and end-to-end test coverage, and ran structured testing and focus groups with dispatchers and drivers.
Naya Labs
Senior Software Engineer
Jan 2022 – Mar 2024
Washington DC, USA (Remote)
  • Forward-deployed with client teams on 3–6-month engagements across the US, UK, Canada, Kenya, Ghana, Nigeria and Pakistan.
  • On OmniArsenal, a US defence contractor's Go CI/CD platform of around eight services behind a React front end, built the ingestion pipeline and the monitoring infrastructure around it.
  • Traced a month-old production defect that broke real-time rendering to an incorrect browser-storage reference.
Sonolabs
Software Engineer
Apr 2021 – Jan 2022
Nairobi, Kenya (Remote)
  • Owned the computer-vision pipeline behind automated meal logging: from one photo of a plated meal, segment each item and classify it by food type, portion, food group and nutrition.
  • Curated and annotated a dataset of locally available Kenyan foods, as no suitable public dataset existed for the cuisine.
  • Built the Flutter nutrition application around it, including a recommender tailored to users' health conditions and goals.
  • Delivered backend features on a Go monolith, integrated the M-Pesa Daraja API, mentored a junior engineer, and introduced unit and integration testing.
Fundle
Lead Software Engineer
Jan 2021 – Apr 2021
Kigali, Rwanda (Remote)
  • Built a financial-literacy search engine in Python and Django with Elasticsearch.
  • Added a TensorFlow natural-language model for voice-based search over a Python and Flask microservice architecture.
  • Mentored the engineering team on Python standards and test-driven development.
Independent
Freelance Software Engineer (Contract)
Aug 2017 – Jan 2022
Nairobi, Kenya (Remote)
  • Designed, built and maintained web and backend systems for startups and small businesses across East Africa and abroad.
  • Cut response times with caching and fewer database round trips.
  • Migrated large systems between AWS and GCP in both directions.

Research & packages

Four packages on PyPI, and four records on Zenodo. All four records are under revision: a software audit of my drug-target platform found defects in its causal-inference stage, the affected quantitative results are withdrawn, and corrected versions of the records are in preparation. The research page states which claims the audit withdrew, which stand, and what has been re-measured since.

CausalPhylo
A published Python library implementing Pearl's backdoor criterion over knowledge graphs, separating edges that can carry a causal claim from those asserting only co-occurrence. Single dependency, typed, documented, on PyPI.
pip install causalphylo
ligase
Reads biological dimensions (organism, reference build, coordinate system, sample order, identifier namespace) out of the files, propagates them, and type-checks every edge when the graph is built, before a step runs. A GRCh37 reference meeting a GRCh38 alignment is a DimensionError, not a plausible wrong number.
pip install ligase-bio
karani
A job-search agent over nine boards that drops roughly 95% of postings on deterministic rules before any model call, and never submits an application.
pip install karani
pict
CLI project scaffolding for Go, TypeScript, Rust and Python, published on PyPI as the fourth package.
pip install pict

Skills

Languages
TypeScript · Python · Go · Rust · SQL
Backend & architecture
Node.js · Express · NestJS · FastAPI · Flask · Gin · gRPC · microservices · event-driven architecture · multi-tenant systems · CQRS · DDD
Payments & fintech
Stripe · Plaid · Modern Treasury · M-Pesa Daraja · ACH · billing systems · payout flows · reconciliation · transaction monitoring · entity resolution
Data
PostgreSQL · MongoDB · MySQL · Redis · Neo4j · Qdrant · Elasticsearch · Snowflake · BigQuery · Splink · Dask · ETL and ELT pipelines · schema design
ML, research & AI
PyTorch · TensorFlow · DoWhy · do-calculus and identifiability · computer vision · dataset curation · LangGraph · multi-agent orchestration · RAG · vector search · recommender systems · anomaly detection
Cloud & DevOps
AWS (ECS, Fargate, Lambda, S3, RDS) · GCP · Docker · Kubernetes · Terraform · CI/CD · GitHub Actions · cloud cost optimisation · observability (Prometheus · Grafana · Sentry · PostHog)
Frontend & mobile
React · Next.js · Flutter · React Native
Practices
Test-driven development · unit, integration and end-to-end testing · code review · technical mentoring · remote and asynchronous collaboration · GDPR, HIPAA and SOX control frameworks

Education

B.Sc. Computer Science
Laikipia University, Kenya — Second Class Honours, Upper Division
Graduated December 2022
Applied Data Science Program
MIT Professional Education — Professional Certificate
January – April 2024