Vishesh Sharma

Senior Software Engineer

San Francisco, California, United States6 yrs 6 mos experience

Most Likely To SwitchAI Enabled

Key Highlights

Expert in building distributed systems at high scale.
Led critical services with 70M QPS and 4 9 5 availability.
Strong focus on AI Agents and Model Interpretability.

Stackforce AI infers this person is a Backend-heavy Fullstack Engineer with expertise in AI and Cloud technologies.

Contact

Skills

Core Skills

Machine LearningDistributed SystemsLarge Language Models (llm)Ai AgentsIdentity And Access Management (iam)Large Scale DevelopmentSoftware Development

Other Skills

CC++Core JavaData StructuresGitHTMLJSpJavaJavaScriptMySQLOrchestrationPlatform as a Service (PAAS)SQLServletsSpring Framework

About

I am a SWE + ML engineer (with a knack for experimentation) experienced in building distributed systems at high scale. I have experience working on research problems. My current work in research focusses on Interpretability and realtime policy enforcement using Agents. I like to read fiction and I have an interest in public speaking. My twitter - https://x.com/vishessh_t YouTube - https://www.youtube.com/@TheJourneyAngle

Experience

Google

3 roles

Senior Software Engineer

Sep 2025 – Present · 6 mos

Gemini Cloud Abuse. SWE+ML role - Building a platform for unsupervised anomaly detection using clustering algorithms for realtime enforcement.

Unsupervised LearningPlatform as a Service (PAAS)Machine LearningDistributed Systems

Senior Software Engineer

Apr 2025 – Oct 2025 · 6 mos

Cloud AI research
1. LLM Deception - MVP on Model Interpretability and In-Context-Learning to steer the models preventing deception.
2. Verified code generation - Contributed to Google ADK for text to policy code generation and policy enforcement adding security as a first
Received offer from Google Research given my work in Cloud AI
Team lead for a highly critical service in the serving path of Google requests (*70M QPS*). Built the service for Gaia ID management serving at this scale with 4 9 5 availability.
Owned an e2e project partnered with multiple orgs that was showcased in Google Cloud Next'25 - App Centric Management -
https://www.forbes.com//2025/04/09/google-cloud-introduces-application-centric-cloud-for-developers/

Large Language Models (LLM)AI AgentsDistributed SystemsPlatform as a Service (PAAS)

SWE III @Google

Nov 2022 – Apr 2025 · 2 yrs 5 mos

Cloud AI Research - LLM Agent safety thru verified code generation.
1. Architected a Secure ADK Plugin (SecureAgent) that connects to VeriGuard Cloud Run service to "compile" natural language security requirements into verified, executable Python policies on startup.
2. Engineered a sandboxed execution environment to mitigate malicious code policy injection and implemented context-aware guardrails to prevent cross-tool data leakage
3. Enabled granular control by injecting security hooks at every stage of the agent lifecycle (pre/post-tool execution and pre/post-model invocation)
4. Implemented a Runtime Tool Wrapper that intercepts specific actions (for e.g. apply_discount), successfully demonstrating multiple flows.
Cloud Security team - Identity and Infrastructure
1. Designed and executed Lifecycle management of GCP resources across all GCP services.
2. Led the project for moving away from the legacy account management system (having Google wide dependency) with clear goals and impact on the OKR to a new distributed ID management service (designed and developed by me) ensuring high reliability.

Large Scale DevelopmentIdentity and Access Management (IAM)Orchestration

Amazon

3 roles

SDE-2

Sep 2022 – Nov 2022 · 2 mos

I worked for the Sellers team. Optimizing the amount Amazon spends in moving the boxes and shipping it safely to the customers.

Software Development

SDE-2

Nov 2021 – Sep 2022 · 10 mos

Project: Config Plane Orchestration
Designed architecture for configuration plane. Designed orchestration of multiple services automating the onboarding process and business configurations.
Created sync and async API's with defined contracts for configuration plane. Onboarded API's to OAuth2.0 and exposed it to orchestrator.
Used CDK to create a generic stack (infrastructure) for supporting any kind of sync or async API's.
Technologies - Lambda, Step functions, API gateway, Python, Dynamodb, S3
Project: Deprecating RDS
Description
Designed architecture for deprecating RDS (and replacing with the new architecture) handling all the dependencies, migrating the data from one schema to other with minimal possible downtime in an incremental fashion ensuring idempotency in the architecture.
Designed Recon system to validate the incremental transfer of data as part of migration.

Software Development

SDE

Aug 2020 – Oct 2021 · 1 yr 2 mos

Building realtime, accurate and scalable solution for hourly associates work schedule, punches and pay.
1. Setup Amazon RDS, schema design and proposed columnar database for analytical queries.
2. Domain Job Optimization:
Worked on redesigning and rewriting a job which stores the domain data into our system. Used S3 instead of RDS in the new design separating compute and storage.
In the new design, instead of streaming the data used batching (after getting alignment on the SLA's) and was able to reduce the job run time by 10X (p99). It also reduced the max age queue from multiple days to couple of hours.
With the new design of batching, was able to scale up the job by 5X because of batching benefits and S3 based architecture. Ensured that batching and compute are done independently.
3. Batch Ingestion:
Designed and implemented batch ingestion architecture to ingest data from 25 different datasets and scaling it to billions of records and TB's of data and vending it at the same time realtime.
Designed recon system for ensuring no DQ issues.
Segregated compute and storage in the new design so that both can scale independently.
Used infrastructure as code to build a Full CD pipeline for datalake.
Also worked on automating QuickSight datasets deployment without manual intervention.
4. Service Migration:
Worked on migrating 6 services from MAWS to NAWS (Native AWS). Ensured security aspects by using latest security protocol OAuth 2.0 for authn/authz
5. Launches:
Worked on launching the product in USA and Morocco. Got all the business requirements clarified and came up with technical solutions for them.
Languages used are Typescript for infrastructure, Python for writing source ingestion code, PySpark API for writing glue scripts written in Python and Java for writing services and jobs.

Software Development

Crio.do

DSA Mentor Crio

Jun 2020 – Jul 2020 · 1 mo

Mentored a group of 20 students. Taught advance DSA and backend concepts. Received excellent feedback.

Software Development

Sharechat

SDE Intern

Jan 2020 – Jun 2020 · 5 mos · Bangalore, India

Led the Karma project for a scale of 14M Daily active users and total user base of 100M to create services and jobs for Karma system.
Karma system encourages users on the basis of their actions by providing Karma points. Karma can be directly associated with the engagement of the user.
Developed robust API's end-to-end with Redis as cache mechanism and Spanner as persistent storage.
Some challenging features implemented on scale:
1. Real time Karma update of users
2. Showing user analytics like engagement of users in real time, Top posts of a user. Generalized this for N days
3. Showing and updating Karma leaderboard in real time
4. Karma as a feature to the recommendation models.
Tech Stack: Node Js, Java
Cloud: GCP
Resources Used: PubSub, Spanner, Redis, BigTable

Software Development

Amazon

Software Engineering Intern

May 2019 – Jul 2019 · 2 mos · Hyderabad Area, India

Designed and implemented Dynamic Report Builder for generating reports dynamically on the basis of user preferences like language, currency, site, role of user and type of report.
Automated report generation process
Technology Used are AWS RDS, S3, Java, Python

Software Development