Vikranth M D profile photo 2 on Tal
Vikranth M D profile photo 1 on Tal
Vikranth M D profile photo 2 on Tal
Vikranth M D profile photo 1 on Tal

Vikranth M D

CSE student with a keen interest in artificial intelligence and machine learning

East Godavari

4+ years of work experience

Quantiphi Analytics Solutions
Feb 20242 yrs 8 mosPresent

On-site

Intelligent Document Processing Pipeline for Insurance | LangChain, RAG, ChromaDB, AWS, GitLab CI/CD • Automated manual document workflows for an entire processing department, accelerating end-to-end processing speed by 90% for 8,000+ documents/month while maintaining a multi-cloud (AWS + GCP) GenAI system across 4 environments via GitLab CI/CD. • Designed cascaded LLM routing (Nova-Lite to Claude 3 Sonnet) with regex-based ACORD detection; implemented per-token cost tracking across model providers and CloudWatch dashboards to alert on model degradation. • Built a RAG validation subsystem (ChromaDB + Sentence Transformers) that flagged out-of-distribution LLM outputs; reduced hallucination rate from 12% to 3% and improved extraction accuracy from 78% to 92%, tracked as live SLA metrics. • Engineered a post-processing validation engine with business-rule enforcement and RapidFuzz-based OCR correction, reducing manual post-editing by 42% and enabling automated downstream ingestion into visualization and downstream systems. • Reduced Kubernetes cold-start latency by 65% (avg 12s → 4.2s) via PyInstaller single-binary Docker builds; implemented structured logging and centralized exception handling, improving throughput by 18%. Automated Gemini Fine-Tuning Pipeline | GCP, Vertex AI, Cloud Run, Cloud Functions, Python, MLOps • Cut Vertex AI compute utilization by 35–40% by implementing distributed locking, dataset validation, and delta-based retraining thresholds, eliminating redundant training runs. • Built an event-driven LLM fine-tuning pipeline on GCP (Vertex AI, Cloud Run Jobs) with automated job status monitoring and self-healing recovery mechanisms for continuous model reliability. • Reduced mean time to recovery (MTTR) for retraining incidents by automating the fine-tuning lifecycle – orchestration, real-time monitoring, metadata tracking, and alerting dashboards – enabling proactive failure detection. • Accelerated production release cycles by documenting
Quantiphi Analytics Solutions
Feb 20247 mosAug 2024

Hybrid

PRs Software Services
Aug 20234 mosNov 2023

India

Nippon Paint India
Aug 20222 mosSep 2022

India

Conducted market research on the Decorative-paints industry Developed marketing strategies for new product launches Analyzed customer data to identify trends and opportunities
Growth Heads Digital
Dec 20201 yr 1 moDec 2021

GitHub
@vickymdn
Medium
@mdvikranth

Vikranth’s AI Stack

claude code
claude code
gemini
gemini
hugging face
hugging face
langgraph
langgraph
openai
openai
LlamaIndex
LlamaIndex

Education

Siksha O Anusandhan Corporate Relation

Bachelor of Technology - BTech, Computer Science

Siksha O Anusandhan Corporate Relation

20204 yrs2024
Siksha 'O'​ Anusandhan University

Bachelor of Technology, Computer Engineering

Siksha 'O'​ Anusandhan University

Join Vikranth and 10k+ other professionals on

Tal Boss

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

Download Tal Boss

Vikranth M D on Tal

Headline

CSE student with a keen interest in artificial intelligence and machine learning

Location

East Godavari

AI stack

claude code

gemini

hugging face

langgraph

openai

LlamaIndex

Work experience

Machine Learning Engineer at Quantiphi Analytics Solutions, Feb 2024 - Present · 2 yrs 8 mos, On-site, Intelligent Document Processing Pipeline for Insurance | LangChain, RAG, ChromaDB, AWS, GitLab CI/CD • Automated manual document workflows for an entire processing department, accelerating end-to-end processing speed by 90% for 8,000+ documents/month while maintaining a multi-cloud (AWS + GCP) GenAI system across 4 environments via GitLab CI/CD. • Designed cascaded LLM routing (Nova-Lite to Claude 3 Sonnet) with regex-based ACORD detection; implemented per-token cost tracking across model providers and CloudWatch dashboards to alert on model degradation. • Built a RAG validation subsystem (ChromaDB + Sentence Transformers) that flagged out-of-distribution LLM outputs; reduced hallucination rate from 12% to 3% and improved extraction accuracy from 78% to 92%, tracked as live SLA metrics. • Engineered a post-processing validation engine with business-rule enforcement and RapidFuzz-based OCR correction, reducing manual post-editing by 42% and enabling automated downstream ingestion into visualization and downstream systems. • Reduced Kubernetes cold-start latency by 65% (avg 12s → 4.2s) via PyInstaller single-binary Docker builds; implemented structured logging and centralized exception handling, improving throughput by 18%. Automated Gemini Fine-Tuning Pipeline | GCP, Vertex AI, Cloud Run, Cloud Functions, Python, MLOps • Cut Vertex AI compute utilization by 35–40% by implementing distributed locking, dataset validation, and delta-based retraining thresholds, eliminating redundant training runs. • Built an event-driven LLM fine-tuning pipeline on GCP (Vertex AI, Cloud Run Jobs) with automated job status monitoring and self-healing recovery mechanisms for continuous model reliability. • Reduced mean time to recovery (MTTR) for retraining incidents by automating the fine-tuning lifecycle – orchestration, real-time monitoring, metadata tracking, and alerting dashboards – enabling proactive failure detection. • Accelerated production release cycles by documenting

Machine Learning Engineer at Quantiphi Analytics Solutions, Feb 2024 - Aug 2024 · 7 mos, Hybrid

Data Science Intern at PRs Software Services, Aug 2023 - Nov 2023 · 4 mos, India

Intern at Nippon Paint India, Aug 2022 - Sep 2022 · 2 mos, India, Conducted market research on the Decorative-paints industry Developed marketing strategies for new product launches Analyzed customer data to identify trends and opportunities

Social Media Marketing Team Lead at Growth Heads Digital, Dec 2020 - Dec 2021 · 1 yr 1 mo

Education

Bachelor of Technology - BTech, Computer Science, Siksha O Anusandhan Corporate Relation

Bachelor of Technology, Computer Engineering, Siksha 'O'​ Anusandhan University

Public links

github: https://github.com/vickymdn

medium: https://medium.com/@mdvikranth