Open to opportunities

Ashutosh Dubey

@ashutoshdubey2

Message

Full-stack developer building GenAI-driven, low-latency apps with scalable architectures.

India

Message

What I'm looking for

I’m looking for a team where I can build scalable full-stack products and ship production GenAI features—especially low-latency RAG and prompt evaluation—while owning performance, reliability, and great UX end to end.

I’m a Computer Science Engineering graduate focused on building full-stack products that feel fast, reliable, and user-centric. I bring strong software engineering fundamentals—frontend architecture, API integration, and database design—so features ship cleanly and scale confidently.

In my GenAI work at Outlier.ai, I crafted and optimized multilingual prompts to improve LLM task success and answer quality. My evaluations showed an 18% increase in task success rates, a 22% reduction in hallucinations, and improvements in consistency through monitoring, drift analysis, and edge-case handling.

Across my projects, I’ve built AI-powered SaaS experiences like an AI meeting intelligence platform with a low-latency RAG pipeline using Gemini and Pinecone, plus asynchronous processing with AWS Lambda and S3. I also developed a smart finance dashboard with Gemini-driven insights and interactive visualizations, prioritizing performance, scalability, and real-world usefulness.

Experience

Work history, roles, and key accomplishments

Generative AI Trainer

Outlier.ai

Feb 2024 - Apr 2025 (1 year 2 months)

Crafted and optimized multilingual prompts (instruction chaining, Chain-of-Thought, Tree-of-Thought) to improve LLM task success and answer quality, increasing task success rates by 18%. Used Evidently AI for data quality monitoring and drift analysis, reducing hallucinations by 22%, ambiguity-induced errors by 30%, and improving model accuracy by 12%.

Instruction Chaining Chain Of Thought LLM Evaluation