Soham Parekh Scandal: Overworking at Multiple AI Firms

he tech industry is grappling with a controversy that has sent shockwaves through the Y Combinator (YC) community and beyond.

It all began on July 2, 2025, when startup founder Suhail Doshi accused software engineer Soham Parekh of working at multiple YC-backed startups simultaneously for over a year, while allegedly failing to fulfill his job responsibilities.

What started as a single post quickly escalated, with other founders corroborating the claims and exposing a pattern of overemployment that has implicated 18 companies.

In California, where most of these startups are based, moonlighting is legally allowed if it doesn’t conflict with primary job duties or intellectual property.

However, Parekh’s case stands out due to the extraordinary number of parallel roles he allegedly managed, raising ethical and operational concerns across the AI startup landscape.

The Resume That Circulated

Parekh’s resume, which has been widely shared online, paints a picture of a highly skilled engineer.

‍

Educated at the Georgia Institute of Technology (M.S. in Computer Science, 2020-2022) and the University of Mumbai (B.E. in Computer Engineering, 2016-2020, with a 9.83/10 GPA), he lists expertise in deep learning, natural language processing, and distributed computing.

His experience includes:

• DynamaoAI (Jan 2024 - Present): Designed interactive dashboards with D3.js and vis libraries, implemented CQRS-based systems using Nest.js, and led AWS Marketplace and SageMaker integrations.
• Union.ai (Jan 2023 - Jan 2024): Developed a workflow orchestration platform with ML and analytic stacks in Go, Python, and React, optimizing KDTree algorithms for distribution and alignment.
• Synthesis (Dec 2021 - Dec 2022): Built a browser-based video editor in React.js with a serverless Flask API, integrating PostgreSQL and MongoDB.
• Alan AI (Jan 2021 - Dec 2021): Created visualization tools and a React.js-based browser environment for training Alan’s AI bot.

This impressive background, combined with a GitHub profile showcasing significant contributions, helped Parekh secure roles at top-tier startups—until the accusations surfaced.

The Podcast ConfessionIn a recent episode of the TBPN podcast, Parekh addressed the allegations head-on. When asked if he worked at multiple jobs simultaneously, he admitted, “It is true.”

He elaborated that financial necessity drove him to work 140 hours a week, a decision he’s not proud of and doesn’t endorse.

Parekh explained that dire circumstances, coupled with his reserved nature, led him to take on multiple roles, not out of greed but survival.

He emphasized his care for the companies, stating he researched their business models and delivered beyond engineering tasks, often sending heartfelt cold emails to secure these opportunities.

When questioned about legal violations, Parekh suggested he believed there might be loopholes, though he didn’t elaborate, focusing instead on his motivations and the quality of his work.

The Companies InvolvedThe fallout has revealed a network of 18 companies that hired or nearly hired Parekh, most of them AI startups founded in recent years. Key examples include:

• Playground AI (YC19): Hired and fired him in 2024 after a week.
• Leaping AI (YC W25): Terminated him upon discovering parallel employment.
• Fuse AI (YC W25): Fired him for working at Leaping AI simultaneously.
• Digger: Onboarded but terminated after background checks, per founder Igor Zalutski, who rated Parekh in the top 0.1% for skills.
• Fleet AI, Lindy, Create, Animetal, Ponder.ai, ComfyUI: All hired and fired him within weeks, citing slow delivery and red flags.

Companies like AIVideo.com, Pally AI, and Mosaic AI interviewed him but declined offers, often due to in-person requirements or post-interview concerns. The YC founder community, led by CEO Garry Tan, played a crucial role in exposing this pattern.

The Deception and Excuses

Parekh, based in India, convinced employers he was relocating to the US, citing an O1 visa process. An email to Fleet AI’s CEO promised an 80% chance of in-person work, while he told Digger’s Zalutski about an apartment in San Francisco’s North Beach.

Evidence suggests he never had US authorization, with one founder alleging a laptop was sent to a fake US address. His excuses—ranging from a military drone damaging his house (debunked by Leaping AI’s Arkadiy Telegin to weather-related internet issues—further fueled suspicions.

The AI Talent Boom and Hiring RisksThe skyrocketing demand for AI engineers, particularly those skilled in large language models (LLMs), explains Parekh’s success. Startups, offering $200K+ salaries, often rushed hires, bypassing thorough vetting.

Parekh’s polished resume and interview performance exploited this urgency, though his “overly green” GitHub activity raised red flags post-hire.

Zalutski admitted Digger’s fast hiring was a calculated risk that backfired.

Implications for Remote Work and Ethics

Remote work enabled Parekh’s overemployment, a trend likely to grow with AI tools boosting productivity.

This raises questions: Can in-person interviews or hybrid models prevent such cases? The debate intensifies as the industry weighs flexibility against accountability.

Ethically, Parekh’s admission of necessity versus intent highlights a gray area—while his work quality was high, the scale of his commitments undermined trust.

Lessons and Looking Ahead

1. Vetting Matters: Background checks and initial in-person assessments could mitigate risks, even for top talent.
2. Community Power: The YC network’s collaboration was key, suggesting broader industry vigilance.
3. Ethical Reflection: Parekh’s case underscores the need to balance personal circumstances with professional integrity.