Abhishek Verma profile photo 1 on Tal

Abhishek Verma

Building LoomStack · Dobr.AI · xAmazon

Bengaluru

7+ years of work experience

LoomStack
May 20262 mosPresent

Remote

We’re building a process orchestration layer for software teams building with AI. Today, humans "manually" leverage AI tools across product specs, planning, coding, testing, and deployment. We invert the workflow: LoomStack owns the end-to-end engineering lifecycle and involves humans only where oversight, judgment, or accountability are needed.
Dobr.AI
Apr 20242 yrs 3 mosPresent

Bengaluru

• Founded Dobr.AI with a vision to create a new standard of technical hiring, to replace the current obsolete interview frameworks, and to democratize evaluation around real-world skills, instead of relying on brand pedigree. • Built a voice-based AI interview agent to conduct natural, conversational technical interviews, simulating human-level interaction and adapting questions based on candidate responses and the ongoing progress on the task. • Architected a scalable, cloud-native microservices platform (AWS) supporting thousands of concurrent interviews with >99.9% uptime. • Built a model-agnostic, agentic orchestration framework using context-driven sub-agents (LLM-based) for question generation, evaluation, and adaptive feedback; integrated LLM routing across Groq, Anthropic, and Bedrock with fallback, caching, and prompt optimization to improve availability, and reduce latency & cost. • Built subsystems for cheating-detection using behavioral signals (typing cadence, fraud signals) and dynamic validation agents. Designed a plugin-based system enabling extensibility for new interview formats (e.g., machine coding, system design) and evaluation logic without core refactoring.
interface.ai
Oct 20258 mosMay 2026

Remote

Technical leader for a 40+ member engineering org. Leading engineering for the core conversational AI platform and the future roadmap at Interface.AI.
Amazon
Feb 20204 yrs 10 mosNov 2024

Bengaluru, Karnataka, India

• Led the development of GenAI workflows, enabling the first-generation GenAI-enabled Kindle device. This involved deploying robust GenAI models, utilizing LLMs and agentic workflow execution, to enhance user experience and content creation on the device. A key aspect of the project was implementing safeguards to prevent the generation of harmful or biased content, aligning with ethical AI guidelines and user safety protocols. This project involved cross-team collaboration, architectural design, planning, and leading a team of 5 SDEs. • Led the end-to-end development of a comprehensive suite of fraud and abuse prevention systems for AWS's self-checkout service. This multi-phased, high-impact initiative directly tackled retail shrinkage (a $94.5B industry loss in 2021). The product suite encompassed core theft detection, alert and monitoring systems, MLOps, and a ground truthing platform. My responsibilities included phased planning, team leadership, cross-functional stakeholder collaboration, and architectural design and execution, leading a team of 4 SDEs. • Led the automated receipt ground-truthing project for Amazon GO, reducing the associate effort required for ground-truthing by 50% and cutting the metric turnaround time by 66% (from 3 days to 1 day). • Led efforts on the automation of camera planning systems for Amazon GO, streamlining the camera planning process by 60% and reducing SDE operational effort by 95%, resulting in increased efficiency and productivity.
Blurbsh Inc.
Mar 20198 mosOct 2019

Remote

• Worked as a founding engineer on the new release of the startup's applications and followup features. • Significantly improved efficiency and performance by using the latest architectural components.

Education

University of Hyderabad

Master of Computer Applications - MCA, Computer Science

University of Hyderabad

Bachelor of Computer Applications, Computer Science

Birla Institute of Technology, Ranchi

Join Abhishek and 10k+ other professionals on

Tal Boss

Tal is the app where bosses meet engineers & builders directly, without middlemen

Download Tal Boss

Abhishek Verma on Tal

Headline

Building LoomStack · Dobr.AI · xAmazon

About

Building the future of software engineering, after 9 years of crafting digital experiences at Amazon and various startups. When I'm not building the next 'awesome,' you'll find me hiking, reading, or playing chess.

Location

Bengaluru

Work experience

Founder and CEO at LoomStack, May 2026 - Present · 2 mos, Remote, We’re building a process orchestration layer for software teams building with AI. Today, humans "manually" leverage AI tools across product specs, planning, coding, testing, and deployment. We invert the workflow: LoomStack owns the end-to-end engineering lifecycle and involves humans only where oversight, judgment, or accountability are needed.

Co-Founder at Dobr.AI, Apr 2024 - Present · 2 yrs 3 mos, Bengaluru, • Founded Dobr.AI with a vision to create a new standard of technical hiring, to replace the current obsolete interview frameworks, and to democratize evaluation around real-world skills, instead of relying on brand pedigree. • Built a voice-based AI interview agent to conduct natural, conversational technical interviews, simulating human-level interaction and adapting questions based on candidate responses and the ongoing progress on the task. • Architected a scalable, cloud-native microservices platform (AWS) supporting thousands of concurrent interviews with >99.9% uptime. • Built a model-agnostic, agentic orchestration framework using context-driven sub-agents (LLM-based) for question generation, evaluation, and adaptive feedback; integrated LLM routing across Groq, Anthropic, and Bedrock with fallback, caching, and prompt optimization to improve availability, and reduce latency & cost. • Built subsystems for cheating-detection using behavioral signals (typing cadence, fraud signals) and dynamic validation agents. Designed a plugin-based system enabling extensibility for new interview formats (e.g., machine coding, system design) and evaluation logic without core refactoring.

Staff Software Engineer - AI at interface.ai, Oct 2025 - May 2026 · 8 mos, Remote, Technical leader for a 40+ member engineering org. Leading engineering for the core conversational AI platform and the future roadmap at Interface.AI.

Tech Lead at Amazon, Feb 2020 - Nov 2024 · 4 yrs 10 mos, Bengaluru, Karnataka, India, • Led the development of GenAI workflows, enabling the first-generation GenAI-enabled Kindle device. This involved deploying robust GenAI models, utilizing LLMs and agentic workflow execution, to enhance user experience and content creation on the device. A key aspect of the project was implementing safeguards to prevent the generation of harmful or biased content, aligning with ethical AI guidelines and user safety protocols. This project involved cross-team collaboration, architectural design, planning, and leading a team of 5 SDEs. • Led the end-to-end development of a comprehensive suite of fraud and abuse prevention systems for AWS's self-checkout service. This multi-phased, high-impact initiative directly tackled retail shrinkage (a $94.5B industry loss in 2021). The product suite encompassed core theft detection, alert and monitoring systems, MLOps, and a ground truthing platform. My responsibilities included phased planning, team leadership, cross-functional stakeholder collaboration, and architectural design and execution, leading a team of 4 SDEs. • Led the automated receipt ground-truthing project for Amazon GO, reducing the associate effort required for ground-truthing by 50% and cutting the metric turnaround time by 66% (from 3 days to 1 day). • Led efforts on the automation of camera planning systems for Amazon GO, streamlining the camera planning process by 60% and reducing SDE operational effort by 95%, resulting in increased efficiency and productivity.

Founding Engineer at Blurbsh Inc., Mar 2019 - Oct 2019 · 8 mos, Remote, • Worked as a founding engineer on the new release of the startup's applications and followup features. • Significantly improved efficiency and performance by using the latest architectural components.

Education

Master of Computer Applications - MCA, Computer Science, University of Hyderabad

Bachelor of Computer Applications, Computer Science, Birla Institute of Technology, Ranchi