CONTACT
--https://trashuaisolutions.framer.ai/
TRASHU
VASHISTH
COMPUTATIONAL LINGUIST &
AI SOLUTIONS ARCHITECT
Chatbot Dev
AI Agents
RAG & Agentic RAG
LLM Integration
Automation
Resume Parsing
2025-Present
LEAD AI DEVELOPER
(FREELANCE / INDEPENDENT)
"Most developers build AI for
Architecting end-to-end AI
machines; I architect AI that
solutions for global clients,
understands human nuance."
focusing on data-sensitive
As a Linguist turned AI Engineer, I
SKILLS
EXPERIENCE
bridge the gap between complex
human language and machine logic. I
don’t just write code; I optimize the
Semantic Core of LLMs. With deep
expertise in Linguistics and RAG
industries (Legal & Finance).
Reducing API costs by 30%
through advanced prompt
optimization and linguistic
compression
2025-Present
(Retrieval-Augmented Generation), I
AI SOLUTIONS DEVELOPER |
FREELANCE / SELF-EMPLOYED
build high-precision systems for Legal,
Legal-Tech Automation: Architected a Legal
Financial, and Recruitment sectors
RAG system that processes 5000+ complex
where 99% accuracy isn't a goal—it's a
requirement
documents using Semantic Search, reducing
lawyers' research time by 80%.
Financial Intelligence Agents: Built
autonomous agents that perform Multi-step
EDUCATION
2023 - 2026
University of Delhi
Bachelor of Linguistics
2025 - 2026
Generative AI Mastermind
Fine Tuning Large Language
Models
SKILLSET
Generative AI: RAG Pipelines &
Autonomous AI Agents.
LLM Frameworks: LangChain, LlamaIndex,
& CrewAI
NLP Tools: spaCy, NLTK, & Hugging Face
Reasoning to analyze P&L statements and
automate financial auditing reports.
Sept 2025- Nov 2025
COMPUTATIONAL LINGUIST & AI
DEVELOPER | PROJECT-BASED
Transformers
Nuanced Sentiment Engine: Developed a sentiment
Programming: Python (Expert), SQL, &
analysis tool for a customer feedback platform that
Bash Scripting.
identifies Sarcasm and Linguistic Nuances which
Vector Databases: Pinecone, ChromaDB,
standard models usually miss.
& Weaviate.
Linguistic Data Strategy: Curated and pre-processed
Models: OpenAI (GPT-4), Claude 3, &
specialized datasets for fine-tuning LLMs, ensuring
Llama 3.
Prompt Engineering
Linguistics: Semantic Analysis, Syntax
Parsing,& Prompt Engineering.
Deployment: FastAPI, Streamlit, & Docker.
high-quality Syntactic and Semantic alignment for
domain-specific tasks.
Zero-Shot Classification: Implemented zero-shot text
classification to categorize thousands of customer
queries into 50+ categories with 90% precision.
