Tanishq Jaiswal — AI Researcher

I’m Tanishq Jaiswal — AI Developer at Samsung R&D. I work on custom-building transformer encoder models like BERT and taking them end to end: data collection → training → architecture tweaks → on-device deployment. That’s the role on paper. What actually keeps me hooked is the infrastructure side of AI. I care about what happens after a model is trained. How it gets optimized, compressed, scaled, and wired into real systems that don’t collapse under production constraints. Lately, I’ve been exploring RAG pipelines, vector databases, memory systems, and multi-agent orchestration using tools like LangGraph and Autogen. Not as shiny demos, but as systems that can survive real workflows. I’m not interested in models that only look impressive in notebooks. I care about shipping AI that works in the wild: lean, reliable, and engineered for scale. 💡 What I build and explore: ~ Encoder model customization (BERT and variants) ~ Efficient fine-tuning and deployment (LoRA, QLoRA, PEFT) ~ Quantization and inference optimization ~ RAG pipelines and memory-driven systems ~ Multi-agent orchestration and infra tooling Always open to connecting with people building AI infrastructure, intelligent agents, and product-first systems.

Stackforce AI infers this person is a highly skilled AI Developer specializing in scalable AI infrastructure and deployment.

Location: Kanpur, Uttar Pradesh, India

Experience: 2 yrs 4 mos