Saheb Motiani
Distributed systems, streaming and ML infrastructure
London, UK · saheb210692@gmail.com · Research · GitHub · LinkedIn
Senior software engineer with 10+ years building production systems at Meta, Babylon Health and Disney Streaming. On an ML-focused career break since 2025, building LLM and RL training systems.
Professional Experience
- Built notifications prompting non-admin users at high-value advertisers to appeal mistaken ad-account enforcements, unblocking $6.6M/year in revenue.
- Ran experiments showing broader appeal notifications increased harm, as bad actors exploited gaps in the appeal process. Worked with data scientists and the abuse-prevention team to limit the rollout to high-value advertisers.
- Fixed a multi-recipient concurrency bug, reducing notification failure rates by 70% for Facebook Pages, 30% for ad accounts and 8% for businesses; found the platform logging false-positive delivery successes.
- Designed a deployment approach using Python stream-processing functions (Faust) for ML risk-prediction models, with outputs consumed by clinical portals and patient-outreach tools.
- Built reusable Flink patterns for ETL, feature engineering and stream processing; the data platform team began rolling out Flink internally.
- Led an 8–10-person cross-functional team delivering a health-scoring product with Applied Scientists; developed Scala services for health-assessment and symptom-checker chatbot workflows.
- Designed distributed services to validate incoming media before transcoding and packaging, connecting on-premises infrastructure and AWS through asynchronous Kafka communication.
Earlier Experience
Software Engineer – Cake Solutions (acquired by Disney) · Nov 2016 – Aug 2017
Built media-acquisition automation for live ESPN+ and MLB.TV production.
Senior Associate – Morgan Stanley · Aug 2014 – Jul 2016
Independent Engineering
ML-focused career break · Aug 2025 – present
- TasteBento: Develop and maintain a live social-discovery app based on users’ taste profiles (Next.js, Postgres/Supabase, Cloudflare), with parallel LLM generation, automated checks and reranking; implemented query parallelization and session-revocation checks for mutations.
- Orbit Wars: Built a GPU-vectorized PyTorch simulator and RL training system (~4.1k end-to-end steps/s on RTX PRO 6000); caught an evaluator counting opponent crashes as wins and added safeguards.
- OLMo GRPO: Built an RL fine-tuning trainer with PyTorch, LoRA and vLLM batched generation.
Technical Skills
Production systems: Python, Scala, Java, SQL, Kafka, Flink, AWS, GCP, Kubernetes; distributed services, streaming pipelines and model deployment.
ML systems: PyTorch, Transformers, PEFT/LoRA, vLLM; training and evaluation.
Education
B.Tech., ICT · Dhirubhai Ambani University (formerly DA-IICT) · 2014
1 / 1