~/ravash · engineer — main · ● ready--:--:--model · 312ms
whoami--bio
Toronto, ON · 15+ years · open to selective collaborations · may 2026
Software engineer. Backend, data, and the plumbing in between.
Ravash Jalil — “RJ”. Fifteen+ years shipping APIs, data pipelines, identity, and the deployment automation that keeps the boring parts boring. Recent work spans network-security data platforms, sports analytics, and a healthy curiosity about agentic systems.
cat./philosophy.md
Software, not showmanship. APIs that don’t flake, pipelines that don’t silently lose rows, deploys that don’t need a human at 2am, and AI systems that are eval-gated rather than vibe-checked — planners, tool loops, and traces with the same engineering rigor as the services they call. The unglamorous parts — tests, CI, observability, schemas, evals — are the parts that decide whether something survives a Monday morning.
— working principle, draft 04
FastAPI · ETL · Auth · CI/CD
Internal platform giving authorized teams secure, audited access to network-security data. FastAPI services on top of Elasticsearch and curated Postgres datasets, an enterprise-identity auth server, and Ansible + GitHub Actions deploys that retired ≈20 hrs/week of manual release work.
FastAPIElasticsearchPostgresAnsibleEntra ID
$24.80 handle+ 4.2%n=148w=σ±.18
w-1w-12w-24w-36now
// pricinglivemulti-season · pandaspₙ → µ
P—02NextWager Pricing Engine2022 — 2025
Sports analytics · ML · Simulation
Player-pricing models built on real-time sports feeds, multi-season pandas pipelines, and a what-if simulation service that sized prize pools and quantified risk before the books opened.
Pythonpandasscikit-learn
a.
amelia@amelia.kfinally shipped the migration. five-week project, zero downtime…
♥ 124 · 18
r.
rohan@rohanrxindexing 2.4M posts now searchable in under 90ms p95
♥ 41 · 6
m.
mira@miradevoauth scopes are finally readable. role inheritance shipped today
♥ 87 · 11
// feed2.4M postsdjango · oauth2elasticsearch
P—03Atrium Network2019 — 2022
Social platform · OAuth2 · Search
Built the platform end-to-end: schemas, posts/likes/subs APIs, OAuth2 role-based access, real-time updates, and Elasticsearch-backed full-text search and discovery.
Django RESTPostgresOAuth2
extract1,247 GB
validate4.81 M rows
transform412 cols
load ▸ pg3.55 / 4.81 M
1,247.3 GB CSV → normalized SQL · airflow dag
// ingestrunningairflow · pandaspostgres
P—04UW Data Ingestion2021
1000+ GB CSV → SQL
Application that parsed, validated, and loaded a terabyte of CSV research data into a structured SQL store. Airflow orchestrating ingestion + transform DAGs; pandas doing the heavy lifting per-shard.
AirflowpandasPostgreSQL
cabpinotsyrahchardrsl'groséport
fruitoaktanninacidsweet
// affinitycosinescikit-learnk=42 · r² 0.81
P—05Wine Recommender2019
ML · Personalization
Custom recommendation system for a wine retailer — personalized picks per customer, designed to lift engagement on the storefront.
Side exploration · LLM tooling
Ongoing tinkering with planners, tool-calling loops, and eval harnesses. Currently a personal sandbox, not production — but it’s where the next platform-grade ideas come from.