Perception · Robotics · Machine Learning

Oltan Sevinc

My domain is perception for the physical world — event cameras, sensor fusion, state estimation. My edge is that I ship it: I took an ML product from research to live on the App Store, solo.

Sydney, Australia Australian citizen — E-3 eligible for US roles Open to perception, robotics & ML roles
Email GitHub LinkedIn Résumé PDF
Portrait of Oltan Sevinc
› PhD candidate — event-based vision & robotics, UNSW › Solo founder — Hango, live on iOS & Android › Teaching — Postgraduate AI, COMP9414
01 · Experience

Shipping production systems

Solo-built, end-to-end, in production.

Founder & ML Engineer · Hango
hango.au · Sydney
Jul 2025 — Present

Built and shipped the entire production stack as the sole engineer — recommender, data pipeline, serverless API, and cross-platform app — live on the iOS App Store and Google Play.

  • Recommender. Personalized feed in PL/pgSQL over pgvector embeddings: interaction-weighted interest clustering, a Bayesian-shrunk trending signal, and explore/exploit balancing — served as materialized candidate feeds with server-side impression tracking, per-cycle deduplication, and archetype cold-start.
  • Group scoring. A group recommendation layer using max-plus aggregation, so one strongly-matched member can carry the recommendation, with a shared-interest breadth factor rewarding broadly shared appeal.
  • Data pipeline. An automated, LLM-powered ETL that scrapes, normalizes, enriches, and semantically deduplicates unstructured event data — 80,000+ raw events → ~8,000 clean templates across 1,600+ venues at an under-1% human-review rate.
  • Serving & infra. API layer as Deno/TypeScript Supabase Edge Functions (search, LLM itinerary builder, push, Places resolution); row-level security across all tables; PostHog and Sentry instrumentation.
  • Frontend. Cross-platform app — iOS, Android, shareable webview — in React Native / Expo using LLM-assisted workflows.
Software Engineer · Honeywell
Intern, retained part-time · Sydney
Dec 2021 — Sep 2022

Joined as a summer intern and retained part-time through the academic year on the backend of Experion, Honeywell's flagship process-control platform.

  • Developed backend features in modern C++ with Boost; automated the nightly build-archiving process in Python.
  • Worked in an Agile team — JIRA, Confluence, Git.
02 · Publications & Research

Event-based vision, robotics & estimation

First-author work · UNSW.

Published · ACRA 2025

Towards Closing the Domain Gap with Event Cameras

M. Oltan Sevinc, Liao Wu, Francisco Cruz

Australasian Conference on Robotics and Automation · first author

End-to-end driving models trained on event-camera data hold their steering performance across day–night lighting shifts far better than grayscale frames — event cameras encode relative brightness change, so the domain gap largely disappears.

Conference
Under review

From Micro-Failures to Macro-Stability: Resolving the Explainability Paradox in Spiking Neural Networks

M. O. Sevinc, L. Wu, F. Cruz

First author · under review · working title

Adapts gradient-based attribution to the spiking domain to make spiking neural networks interpretable, reconciling unstable per-spike behaviour with stable network-level explanations.

Honours Thesis · 2023

Robotic Teleoperation with Haptic Feedback for Remote Ultrasounds

B.E. Mechatronic Engineering (Honours) · supervised by Liao Wu

UNSW Sydney

A real-time haptic teleoperation interface between a Universal Robots UR5e arm and a 3D Systems Touch device over ROS / MoveIt — quaternion-derived angular velocity, deadband + consecutive-zero filtering, and force feedback from the arm's built-in sensor for remote sonography.

Demo video
03 · Education

UNSW Sydney

Computer science & mechatronics.

PhD, Computer Science
Feb 2024 — Present

Thesis: Applications of Spiking Neural Networks in Robotics. Australian Government Research Training Program (RTP) Scholarship.

B.E. Mechatronic Engineering (Honours) & Computer Science (AI)
2018 — 2023

Teaching Assistant since 2022 — Postgraduate AI (COMP9414), robotics, autonomous systems, computer networks; tutorials on EKF, sensor fusion, computer vision, and kinematics.

04 · Technical Skills

The stack

Research focus: event-based vision · sensor fusion & state estimation · spiking neural networks · interpretability / XAI.

Languages
C++PythonSQL · PL/pgSQLTypeScript / JSMATLAB
Robotics & Perception
ROS / ROS2MoveItGazeboExtended Kalman Filterssensor fusionstate estimationevent-based visionOpenCVLIDARkinematics
ML & Data
PyTorch · DDPSNNs · SpikingJellyrecommender systemsvector search / pgvectorLLMs & LLM ETLscikit-learn
Systems & Infra
PostgreSQL / pgvectorSupabaseDeno edge functionsETL pipelinesGCPDockerGitLinuxPostHogSentry
05 · Writing

Notes, in progress

Working writeups on the systems and research above.

Recommenders

Designing Hango's recommender in Postgres

Embeddings, Bayesian-shrunk trending, and archetype cold-start — building a personalized feed entirely in PL/pgSQL over pgvector.

Drafting
Perception

Event cameras vs. the day–night domain gap

Why a sensor that only sees change shrugs off lighting shifts that wreck frame-based models — and what that buys autonomous systems.

Planned
Estimation

Sensor-fusion notes: what an EKF actually buys you

State estimation intuitions from mechatronics that keep paying off in perception and ML systems.

Planned

→ Read the blog for all notes as they're published.