~/$ cat about.md
ABOUT
Building AI products until they learn to build themselves.
Engineering Lead — Applied AI at Deriv (Dubai). 9+ years, four of them inside Target's ML Platform — feature store, Kernels-as-a-Service notebook platform, MLflow, and model delivery. 3× AWS certified. Educator. Professional introvert who thrives on writing to learn and questioning existing processes. Into RAG, MLOps, and C-level dashboards that actually make sense.
$ git log --oneline
- Deriv — Engineering Lead, Applied AI (current)
- Target — Senior → Lead ML Platform Engineer
- Quantiphi — ML Engineer
- Amazon — Business Analyst
- UChicago Graham School — PGP Data Science & ML
$ cd target/ml-platform && ls
4 years · ~400 DS · hybrid on-prem + GCPFour product surfaces over one substrate. Pick a pillar to see the scope.
In-house feature management for ~400 DS
6 weeks → 1 week onboardingRebuilt the Feature Management SDK and onboarding flow so a Data Scientist could declare a feature in one call and ship to production in a week — without skipping a single governance gate.
- Collapsed
create_featuresetto a single contract covering source, entity keys, timestamps, freshness SLA, PII tier, retention. - Converted the call to async Shepherd-orchestrated pipelines across Data Placement, Feature Management, Governance, and Observability APIs.
- Inline Great Expectations checks short-circuited bad features with concrete, actionable errors.
- Emitted lineage events to Lasso — substrate for auto-deprecation and downstream governance.
$ ls certs/
3× AWS (Solutions Architect, ML Specialty, Data Analytics), UChicago PGP Data Science & Machine Learning.
$ echo $TECH
$ cat extracurriculars.md
Speaking, workshops, and beyond.
Wed, Jan 21st · 3:00 PM CET / 6:00 PM GST
Spoke at the lablab.ai × Deriv event on autonomous AI systems. @lablabai & @Derivdotcom on X and LinkedIn.
Want to talk embeddings? Drop a line.