Aniruddha Shrikhande
AI Systems Engineer specializing in scalable Agentic Architectures, low-latency RAG systems, and production LLM infrastructure — serving 200K+ users.
Building the
intelligent infrastructure
of tomorrow.
I design and deploy distributed AI systems that don't just work — they scale. From multi-agent orchestration to sub-second retrieval pipelines, I obsess over the intersection of performance and reliability.
My work spans enterprise clients like AMD, Tata Elxsi, and EXL, where I've shipped production systems handling real users, real data, and real stakes.
Where I've
made impact.
Associate AI Specialist
Association of Data Scientists (ADaSci)
AMD
Developer Engagement Platform- Architected gamification engine with 13 activity types and multi-layer deduplication — powering real-time leaderboards.
- Designed admin analytics dashboard with precomputed PostgreSQL views, enabling sub-second insights across 100K users.
- Engineered serverless social media verification pipeline using Edge Functions — eliminating 90% of manual review.
Tata Elxsi
AI Coding Evaluation Platform- Engineered custom boilerplate generation engine supporting 5+ languages that auto-scaffolds ready-to-code environments.
- Designed seniority-banded AI evaluation framework with deterministic scoring rubrics and validation checks.
EXL
Gamified Learning Platform- Designed and automated generation of high-fidelity synthetic datasets across finance, healthcare, and retail.
- Built scalable data pipelines reducing manual dataset preparation by ~70%.
Internal Projects
ADaSci R&D- Architected real-time AI video interview system with streaming avatar responses, async FastAPI, and resilient S3 uploads.
- Authored 70+ tutorials and co-developed Certified Agentic AI Certification — 50K+ learners impacted.
Selected
work.
Multimodal Clinical
AI Assistant
Built a multimodal AI assistant using DeepSeek and Janus for text, PDF, image, and audio inputs. Architected a low-latency RAG pipeline with hybrid retrieval and evaluation guardrails.
Supply Chain Multi-Agent Collaboration System
AI / Logistics- Developed an AI-driven system enabling autonomous agents to collaborate in real-time, optimizing delivery routes and schedules.
- Employed DeepSeek R1 Distill Qwen 1.5B model for processing and analyzing complex supply chain data.
- Utilized CrewAI framework to orchestrate autonomous agents, each with specific roles and goals.
- Integrated LlamaIndex to manage and query large datasets efficiently.
AI-Powered Financial News Research Tool
AI / FinTech- Engineered financial research system processing 100s of articles, improving research speed by 50%.
- Implemented efficient retrieval pipeline using OpenAI Embeddings with FAISS IVF indexing (4096 clusters).
- Built parallel content extraction handling 100+ URLs with fault tolerance and rate limiting.
- Optimized cost using Redis caching and batch processing, reducing API calls by 35%.
Color & Texture Based Similarity Search to Enable Secure Image Sharing on Social Media Platforms
2023 3rd Asian Conference on Innovation in Technology (ASIANCON)
IEEE