Sabari Vadivelan S.-AI Engineer | LLM Evaluation & RAG Systems | Agentic AI
Check rate
Experience
AI/ML Freelance Contributor
Handshake AI
- Benchmark Design & Adversarial Testing: Design and author benchmark tasks that evaluate frontier AI agents' terminal-based coding and reasoning capabilities for an AI evaluation platform, engineered to survive a multi-stage adversarial review process while every grading rule stays fully disclosed.
- Technical Reasoning: Contribute tasks across a wide range of technical domains, each requiring deep technical reasoning rather than pattern matching.
- AI Evaluation: Serve as an AI Evaluator, assessing AI-generated outputs against task-specific criteria—correctness, relevance, instruction adherence, technical accuracy, and reasoning quality—across coding, technical reasoning, and multi-turn domains.
AI Software Engineering Intern
Virtusa Corporation
- Microservices Architecture: Engineered the "Observe" AI proctoring and examination platform spanning 8 services: React/Vite web client, PyQt6 Windows desktop proctoring app, FastAPI backend, isolated Docker code-execution sandbox, LangGraph-based adaptive/multi-agent question engine, and a report generation/rendering pipeline.
- Production LLM Integration: Built robust FastAPI services connecting application workflows to Groq and Gemini foundation models, optimized for reliable, low-latency LLM integration.
- Security & Telemetry: Implemented a secure Windows proctoring client with runtime hardening (anti-tamper checks) and HMAC-signed telemetry upload; enforced a cross-service security model (JWT for web, HMAC for high-trust desktop operations).
- Observability: Implemented OpenTelemetry tracing to monitor agent decision paths, API latency, model usage, and fallback reliability across the assessment infrastructure.
CyberSLM - Cybersecurity Language Model
- End-to-End LLM Pipeline: Architected a complete pipeline for a 33.5M-parameter decoder-only LLM specialized in cybersecurity: custom BPE tokenizer, large-scale corpus preparation, pretraining, instruction tuning, and benchmark evaluation workflows.
- Model Optimization: Diagnosed and fixed 19 pipeline defects, cutting perplexity by 23% and validation loss from 2.63 to 2.36; published weights on Hugging Face with full before/after methodology.
linkedin-mcp-bridge
- Agentic Browser Automation: Engineered and published an MCP server (npm) giving AI assistants 102 tools across 15 categories to operate LinkedIn through an authenticated browser session—profile management, posts, messaging, job search, recruiter outreach, and more.
- Resilient Architecture: Designed a resilient dual-transport architecture (Voyager API for reads, Playwright for writes) with selector fallback chains and self-healing selectors that use MCP model sampling to regenerate broken selectors from live DOM structure.
ALEC - Autonomous Legal Entity for Contracts
- Risk Assessment: Developed a domain-specific RAG and agentic system for contract analysis, automated clause extraction, and risk assessment, with SHAP-based Explainable AI ensuring generated risk scores are transparent and auditable.
Advanced Multidomain RAG System
- Enterprise Retrieval: Built a scalable hybrid retrieval system combining FAISS semantic search with BM25 lexical search for complex multi-format document QA, with domain routing, table extraction, and explainable output (confidence scores, decision trace). (HackRx 6.0, Top 35 of 46,000+ participants).
Applied Machine Learning Intern
Edunet Foundation | Shell
- Predictive Modeling: Developed end-to-end machine learning workflows in Python for predictive analysis and risk assessment using complex clinical indicators.
- Data Pipeline Engineering: Built data ingestion, cleaning, preprocessing, and feature-scaling pipelines using Pandas, NumPy, and Scikit-learn to prepare datasets for model development.
Crowding and Attribute Extraction System
- Vision-Language Integration: Combined computer-vision models (YOLOv11, DeepSORT, CSRNet) for real-time crowd density tracking and movement monitoring, integrated with Llama-based vision-language processing to extract structured metadata from raw visual/textual inputs. (CodeIQ Hackathon, Winner).
OpenBookAI
- Benchmarking System: Architected a centralized Python backend for comparing outputs from multiple foundation models (ChatGPT, Gemini, Copilot) across prompt variations, with a strict focus on response accuracy, output quality, and latency benchmarking.
Smart Vision Product Counting System
- Large-Scale Vision Analytics: Independently curated, cleaned, and manually annotated a dataset of 70,000+ images; deployed computer-vision workflows for automated object counting, brand recognition, and retail inventory analysis, as team lead. (Flipkart Grid 6.0, Semi-Finalist).
Industry experience
See where this freelancer has spent most of their professional time.
Experienced in Information Technology, Healthcare, and Retail.
Business area experience
See which departments and functions this freelancer has contributed to most.
Experienced in Information Technology, Product Development, Legal, and Business Intelligence.
Summary
AI/ML Software Engineer with hands-on experience in LLM applications, AI evaluation, agentic systems, Retrieval-Augmented Generation (RAG), computer vision, and Python backend development. Experienced in building multi-agent workflows with LangGraph, integrating foundation models, evaluating model outputs, developing FastAPI services, and shipping open-source developer tooling. Strong foundation in Python, machine learning, NLP, prompt engineering, dataset engineering, model evaluation, and production-oriented AI systems.
Skills
- Programming: Python, Typescript, C, C++, Sql.
- Ai / Ml: Machine Learning, Deep Learning, Nlp, Computer Vision, Llms, Generative Ai, Model Evaluation.
- Llm Systems: Prompt Engineering, Ai Evaluation, Rag, Agentic Ai, Langgraph, Langchain, Llm Apis, Instruction Tuning, Benchmark Design, Adversarial Testing.
- Data / Retrieval: Pandas, Numpy, Scikit-Learn, Faiss, Pinecone, Bm25, Semantic Search, Dataset Engineering, Vector Search.
- Backend / Infra: Fastapi, Rest Apis, Docker, Microservices Architecture, Git, Postgresql, Mongodb, Supabase.
- Computer Vision: Opencv, Yolo, Yolov8/V11, Deepsort, Csrnet, Roboflow.
- Security & Tools: Jwt Authentication, Hmac Request Signing, Api Security, Model Context Protocol (Mcp), Playwright, Zod, Node.Js/Typescript Tooling.
Languages
Education
Manakula Vinayagar Institute of Technology
Bachelor of Technology · Computer Science and Engineering · Puducherry, India · 8.67 CGPA
Certifications & licenses
AWS Foundations of Prompt Engineering
AWS
Certiport IT Specialist: Python
Certiport
IBM Data Science
IBM
Oracle Certified Associate: AI Foundations
Oracle
Oracle Certified Professional: Generative AI
Oracle
Statistics
Experience
Global experience
Expertise
Qualifications
Profile
Frequently asked questions
Have questions? Find more information here.
Sabari Vadivelan is based in Singirikudi, India and prefers 100% remote projects.
Sabari Vadivelan speaks the following languages: Tamil (Native), English (Advanced).
Sabari Vadivelan has at least 3 years of experience. During this time, Sabari Vadivelan has worked in at least 3 different roles and for 10 different companies. The average length of individual experience is 3 months. Note that Sabari Vadivelan may not have shared all experience and actually has more experience.
Based on recent experience, Sabari Vadivelan would be well-suited for roles such as: AI/ML Freelance Contributor, AI Software Engineering Intern, Applied Machine Learning Intern.
Sabari Vadivelan's most recent position is AI/ML Freelance Contributor at Handshake AI.
In recent years, Sabari Vadivelan has worked for Handshake AI, Virtusa Corporation, CyberSLM - Cybersecurity Language Model, linkedin-mcp-bridge, and ALEC - Autonomous Legal Entity for Contracts.
Sabari Vadivelan is most experienced in industries like Information Technology, Healthcare, and Retail.
Sabari Vadivelan is most experienced in business areas like Information Technology, Product Development, and Legal. Sabari Vadivelan also has some experience in Business Intelligence.
Sabari Vadivelan holds a Bachelor in Computer Science and Engineering from Manakula Vinayagar Institute of Technology.
Sabari Vadivelan has 5 certificates. Among them, these include: AWS Foundations of Prompt Engineering, Certiport IT Specialist: Python, and IBM Data Science.
Sabari Vadivelan is immediately available full-time for suitable projects.
Daily rate distribution
The rates shown represent the typical market range for freelancers in this position based on recent contracts on our platform.
Average rates for similar positions
Rates are based on recent contracts and do not include FRATCH margin.
The average daily rate is the mean of all daily rates from recent contracts of comparable freelancers on our platform.
The median daily rate is the middle value of all daily rates — half of comparable freelancers charge less, half charge more. Unlike the average, it is barely affected by outliers.
Calculated based on our freelancers’ daily rates as of 23 Sep 2026. Actual rates may vary depending on seniority level, experience, skill specialization, project complexity, and engagement length.
Similar freelancers
Discover other experts with similar qualifications and experience
Noushiq M.
Marcel S.
Manuel P.
Stefan O.
Mathias W.
Stylianos I.
Aruldass A.
Samuel K.
Mirza K.
Rutger B.
Experts recently working on similar projects
Freelancers with hands-on experience in comparable project as a AI/ML Freelance Contributor
Robin L.
Majid K.
Riccardo A.
Antony G.
Michael T.
Lidija V.
Hakan A.
Ilya A.
David P.
Anita B.
Nearby freelancers
Professionals working in or nearby Singirikudi, India
Dinmukhabet A.
Hikmat S.
Javid K.
Sergey F.
Emin M.
Elena G.
Beka C.
Konstantin F.
Giorgi T.
Revaz T.
