Hello, World!  I'm

Apurva Patkeshwar

Tech Lead Machine Learning Engineer

About Me
About Me

Tech Lead at Uber. Previously seen at Microsoft, Amazon, Twitter & Yahoo. NYU alumnus. Future martian astronaut.

Profile Picture

I build large-scale ML and platform systems with a product mindset. My work spans real-time authorization, model serving, latency-sensitive services, and security infrastructure. I care about clarity, reliability, and UX that makes sophisticated systems feel simple. I enjoy the craft of engineering, especially when strong systems thinking meets good design.

Profile

Born and raised in Mumbai, India. Fitness enthusiast. Big-time foodie. Builder at heart ❤

  • Name Apurva Patkeshwar
  • Role Software Engineer
  • Email apurva.p@hotmail.com
  • Location New York, NY

Skills

Strong opinions, loosely held. Shipped in production. Spaces over tabs.

  • 95%
    Technical leadership
  • 90%
    ML Systems
  • 95%
    Shipping under pressure
  • 85%
    Security instincts
  • 80%
    Product thinking
Background

A few highlights from the last decade.

I started at Microsoft in 2014, earned my master’s at NYU, then worked at Yahoo, Amazon, Twitter, and now Uber. The common thread has been building production systems where scale, reliability, and good judgment matter.

Experience

Software Engineer

Feb. 2023 - Present

Uber

Led multi-year access control transformation from static policies to adaptive ML-driven authorization, spanning risk visibility, AI-assisted human-in-the-loop decisions, and ML-based enforcement for agentic workflows.

ML Engineer

Jul. 2021 - Jan. 2023

Twitter

Delivered a fully managed, config-driven ML serving platform that enabled teams across Twitter to serve ML models at scale with 20 ms p99 latency. Built model observability and drift monitoring that surfaced regressions earlier and reduced average retraining frequency while preserving model quality.

SDE

Jun. 2018 - Jul. 2021

Amazon

As a dev on the Automatic Speech Recognition (ASR) team at Alexa, owned a tier-1 web service platform which delivers the fastest, most reliable, and affordable service for audio processing and speech to text conversion. Decoupled Alexa’s audio streaming path from core ASR architecture and moved processing closer to customers, reducing p99 user-perceived latency by 25% in emerging markets.

Software Development Intern

May 2017 - Aug. 2017

Yahoo

As a part of the Advertising & Search Products team, reduced forecasting time from over 7 hours to less than 2 hours and improved search performance with in-memory indexing.

Software Development Engineer

Jul. 2014 - Jul. 2016

Microsoft

Selected for Microsoft Academy for College Hires (MACH), a 2-year new grad development program. Built an office-space web app visualizing real-time occupancy, noise, and temperature data using custom 3D-printed sensors. Added RFID-based room reservation and helped drive a 3x increase in monthly usage.

Software Engineer Intern

Jun. 2013 - Aug. 2013

Microsoft

Implemented a utility for pre performance testing environment setup of Microsoft's Volume Licensing Reporting portal. Also contributed to building a stable and efficient automation framework and automated regression test cases which drastically reduced manual testing effort for every consequent release cycle.

Education

Master of Science

Sep. 2016 - May 2018

New York University

Courant Institute of Mathematical Sciences

Major: Computer Science
GPA: 3.7/4.0

Bachelor of Engineering

Aug. 2010 - Jun. 2014

University of Mumbai

Sardar Patel Institute of Technology

Major: Information Technology
GPA: 3.6/4.0

Selected Work

A few systems and products I helped ship.

Testimonials

What past managers said about working with me.

First and foremost, Apurva is a delight to work with. He is passionate and eager to help out broadly in the team, bringing a bright smile and enthusiasm to every call and conversation. In our (unfortunately) short tenure together at Twitter, I've seen Apurva shine as a mentor, set up monitoring/alerting for core services, work with Product to prototype and test hypotheses, and help mature the team's processes to be more efficient and effective.

Sean Paradiso

Manager (Machine Learning), Twitter

Apurva was part of the ML Serving team that focuses on ML inference workloads. Apurva has been a key team member that brought a lot of team cohesion and energy through scheduling social meetings etc. This has been important to that team as the team was distributed across multiple locations.

Briac Marcatté

Senior Manager (ML Eng), Twitter

Coming straight from college, he was able to quickly ramp up in the domain and the technology being used with minimal guidance. The innovative ideas Apurva implemented proved to be very beneficial and helped the team meet business expectations of delivering in a short duration of time and with zero defects. Along with focusing on his assigned project work, he proactively volunteered for stretch assignments. Overall, Apurva has consistently performed his duties with competence and gone the extra mile to ensure a solid product was shipped. He gelled very well with the team and was respected by his team members for his passion and innovative ideas.

Rajesh Chadha

Sr. Software Engineering Lead, Microsoft

Apurva showed tremendous enthusiasm for technology and played a crucial role in some of the automation delivery for my team. He is a self starter and very committed team player. I have never known a situation when either I or team leads had to follow up with him. He would go the extra mile to scale up on new technology and coding skills. Even though he was still a student while doing internship, at workplace he displayed high levels of maturity and great sense of responsibility. I was happy he was in my team.

Ronnie Joseph

Senior Manager, Microsoft

10

Years Shipping Production Systems

2

Countries Worked

4

Big Tech Companies

2

Academic Degrees

6767

Lines of Curiosity

INT_MAX

Tokens Used
Contact

Let’s build something useful.

Whether it is ML systems, security infrastructure, or platform tools, I’m always happy to compare notes.