Joseph JacobAnjilimoottil
Full-stack engineer in Dublin. I own products end to end — Django REST APIs, async data pipelines, React frontends, and AWS-backed ML systems — shipped, instrumented, and scaled.
One engineer,
the whole stack.
At Easebuzz I didn't hand work off between teams — I shipped each layer myself, from the pixel to the pipeline. This is the shape of a typical system I owned in production.
React + TypeScript dashboards, merchant self-serve onboarding flows, internal tooling UIs.
Django REST infrastructure serving 50,000+ partners — versioned, documented, zero major incidents.
Async Python decision pipelines that cut backend load 70% across 150,000+ daily transactions.
Transactional stores, caches and streaming aggregates — 8,000+ merchants onboarded at scale.
Docker + Jenkins CI/CD across 10 services — deploys from 2 days to 15 minutes, multi-AZ on AWS.
Joined Easebuzz and built the Partner Program platform end to end.
Async commission engine, mass onboarding API, CI/CD across 10 services.
Promoted twice in 22 months. Led partner-facing products and the first in-house AI support chat.
At Easebuzz I owned full-stack products behind a payments platform handling 150,000+ daily transactions — from React self-serve dashboards and Django REST APIs to async decision pipelines in Python asyncio and AWS-deployed ML services. The work is measurable: faster, cheaper, and self-serve.
Backend processing load cut across 150,000+ daily transactions by replacing synchronous logic with async pipelines.
A 3-day multi-approval onboarding process collapsed to 30 minutes for 8,000+ merchants.
Reusable Django REST infrastructure serving 50,000+ partners with zero major incidents post-launch.
Mass merchant onboarding scaled to 8,000+ merchants and contributed 30% gross revenue growth.
Deployment time from 2 days to 15 minutes across 10 services with Docker and Jenkins CI/CD pipelines.
Delivered the company's first AI-powered customer support chat on Amazon Bedrock, trained on internal product policy and deployed across multiple AWS availability zones.
Employee of the Quarter within 3 months of joining · promoted twice in 22 months · Rank 2, SIT CodeChef “The Bug Detective”.
A transformer for early cognitive signal
Extended an Interpretable Multi-IT Transformer with Triplet Attention to stage Alzheimer's across 4 classes on 44,000 MRI scans, optimising for clinically verifiable regions rather than benchmark scores alone. Trained on an NVIDIA A100 with mixed precision and cosine annealing, benchmarked against 4 CNN baselines — showing transformers need multimodal or volumetric input to match CNNs on 2D-MRI-only tasks.
Moby Dublin — Spark telemetry
An event-time-windowed Spark Structured Streaming pipeline with fault-tolerant aggregation for a distributed fleet telemetry system, surfacing demand hotspots and crowding signals across AWS S3 (Parquet) and MongoDB.
Skin cancer classification
CNN models for skin lesion classification in TensorFlow and Keras — InceptionV3, MobileNetV2 and VGG variants with augmentation, reaching 86% accuracy.
View project →Data mining & ML
End-to-end analysis: cleaning, feature engineering and EDA, then supervised models in scikit-learn compared across evaluation metrics.
View project →Akshaya Patra fleet tracker
Intern project tracking food vans dispatched from kitchens to schools — driver and route details, built in Java with Git.
View project →Notes fromthe build
Short write-ups on the systems I build — pipelines, payments infrastructure and models — and what I would do differently next time.
Dublin · Jan 2025 — Jan 2026
Pune · Jul 2019 — May 2023 · Honours in Artificial Intelligence and Machine Learning
Transmit a signal.
Open to full-stack, backend, data and ML engineering roles in Dublin and remote.