Mar 2025 – Present · 1 yr 2 mos · San Francisco, CA

• I help engineers build and accelerate their careers in AI/ML engineering.
• I've spent 25 years in engineering, building super scale AI/ML training infrastructure and Instagram Ads AI/ML systems at Meta, and ML platforms at Yahoo. Over that time, I noticed a pattern: brilliant engineers with real depth were stalling in their careers, not because of skill gaps, but because they lacked the right guidance to channel their ability into real growth.
• www.TechnicalRound.com exists to fix that. I offer 1:1 coaching sessions where we dig into real system design thinking, ML fundamentals, and the strategic clarity you need to grow or transition into AI/ML engineering roles at any level.
• What makes this different from generic coaching:
• → Sessions are led by someone who's actually built and shipped AI infrastructure at scale
• → We work on real, messy, ambiguous problems — the kind you'll actually face on the job, not in a textbook
• → You get a clear picture of where you stand and a concrete plan for where to go next
• → Direct, honest feedback — the kind your friends won't give you
• If you're serious about building a career in AI/ML engineering and want a guide who's been in the trenches, book a session at www.technicalround.com

Dec 2019 – Feb 2025 · 5 yrs 2 mos · Menlo Park, CA

• Led multiple AI Infrastructure teams of 20-30 engineers and scientists, including 2-3 engineering managers and principal engineers, building the core systems behind Meta's AI/ML capabilities across Facebook, Instagram, WhatsApp, and Ads.
• Scope:
• Owned the AI infrastructure stack including training infrastructure, model evaluation pipelines, inference systems, checkpointing, and data preprocessing. These systems served 5,000+ ML engineers across Meta and supported thousands of training jobs daily on a 14,000 GPU cluster.
• Key impact:
• Led Instagram Ads delivery teams to a 15%+ increase in Ads revenue, directly contributing to company earnings.
• My team's infrastructure and tooling were used in the development of Meta's 3D parallelism partitioning architecture for LLaMA model training. I contributed to key technical decisions.
• Built and scaled evaluation systems for offline and inference evaluation of models with billions of parameters, supporting product teams across Ads, Recommendations, Instagram, WhatsApp, and Integrity.
• Partnered with senior cross-functional leaders across General AI, Ads Ranking, and Recommendation Algorithms to define AI infrastructure strategy and roadmap.
• Represented the AI Infrastructure department at Meta-wide engineering events, communicating vision and technical direction to broader engineering leadership.

Sep 2011 – Dec 2019 · 8 yrs 3 mos · San Francisco Bay Area

• Led 10-20 engineers and data scientists, including engineering managers, building AI/ML-powered Trust and Safety systems that protected Yahoo's entire advertising ecosystem.
• Key impact:
• Led teams that developed AI/ML systems scanning millions of Ad creatives and billions of supply-side URLs daily, preventing fraud and protecting advertisers at scale.
• Built the ML Platform team from scratch (0 to 6 engineers across multiple geographies), creating foundational frameworks and tooling adopted across Yahoo DSP, Yahoo Ad Exchange, Yahoo Native Ads/Gemini, and Panama.
• Evolved the platform architecture over 8 years from batch processing (Bigdata, Hadoop, Pig, Oracle) to real-time streaming (Kafka, Storm, Spark), while transitioning ML models from classical techniques to Deep Learning.
• Developed GDPR compliance scanning, political ad detection, and brand impersonation detection systems.

Mar 2010 – Sep 2011 · 1 yr 6 mos · Shenzhen, China and Bangalore, India

• Led architecture for Huawei's Big Data platform and Cloud Platform.
• Developed high-availability solutions for critical Hadoop components and built cloud infrastructure enabling auto-scaling of enterprise applications.

Jun 2007 – Feb 2010 · 2 yrs 8 mos · London, United Kingdom

• Designed and built enterprise VoIP servers and a cloud-based communication service from scratch, opening new revenue streams for the company.
• Developed a log analysis and debugging platform similar to Splunk for parsing large-scale operational data.

Oct 2005 – May 2007 · 1 yr 7 mos · Bengaluru Area, India

• Founding member. Scaled the company from 3 employees to 100+ team members and 35,000+ paying customers.
• Led technology strategy and product development.
• Successful exit through acquisition.

Jul 2000 – Oct 2005 · 5 yrs 3 mos · Bangalore Urban, Karnataka, India · On-site

• Motorola Solutions (Kodiak Networks), Ness Technologies (acquired by Oracle), LogicaCMG, Subex.
• Roles spanning Senior Engineer to Senior Systems Analyst.
• Built telecom billing systems, fraud detection products, real-time settlement systems, and push-to-talk communication platforms.