Shasank Polamraju

With over a decade as an AI Engineer, my ambition and deep curiosity drive me to continuously explore complex challenges, always seeking mastery in developing innovative AI solutions. I approach my work with determination, applying insightful and empathetic perspectives to create impactful technologies.

Experience

A

Al Engineer

Analytos Al · Jan 2026 – Present

Not yet confirmed

• Partner directly with CEO, CTO, and enterprise clients to define Al system architecture, conduct technical demonstrations, and deliver scalable Al solutions.

Artificial Intelligence Engineer

analytos.ai · Jan 2026 – Present

Not yet confirmed

• Partnering directly with the CEO, CTO, and enterprise clients to translate complex requirements into production-grade Agentic AI systems and multi-agent platforms.
• 🚀 Key Enterprise Deployments
• 1.
• Cosellus — Distributed Multi-Agent Orchestration & Graph-RAG
• Architecture: Engineered a deterministic 11-step multi-agent platform using LangGraph to automate complex, asynchronous enterprise logic.
• State & Concurrency: Integrated Celery and Redis for distributed task routing with state checkpointing, workflow replayability, and PostgreSQL JSONB persistence.
• Graph-RAG & Streaming: Linked Neo4j with ChromaDB for citation-backed retrieval; built real-time streaming agent logs via Redis Pub/Sub and WebSockets.
• 2.
• Inspectly — Multimodal Engineering Drawing Intelligence
• VLM Optimization: Designed a tile-based Gemini Vision pipeline to bypass fixed-token limits on high-res GD&T sheets, doubling dimensional recall ($49 \rightarrow 96$ callouts).
• Hybrid Framework: Isolated VLM failure modes via ablations; architected an OCR-anchored hybrid model (OCR for localization, LLM for interpretation) to eliminate hallucinations.
• RAG Recovery: Fixed context propagation bugs in a LangGraph multi-agent pipeline over ASME standards, restoring grounding accuracy from 2/41 back to 40/41 rows.
• 3.
• InvoiceIQ — AI-Powered Automated Invoice-Coding Platform
• Core Automation: Built a multi-tenant AP automation platform (FastAPI, LangGraph, React, AWS) mapping invoice line items to NetSuite GL accounts.
• Deterministic Engine: Built a no-code rule compiler converting English to Python at authoring time; executed runtime inside sandboxed RestrictedPython for compliance.
• Scale & DevSecOps: Integrated a versioned DB with a validation gate, scaling onboarding to 400+ vendors.
• Deployed via keyless AWS ECS Fargate + OIDC across isolated Dev/Prod.

A

AI Product Engineer

AI Technology Venture (Self-Employed) · Jul 2025 – Dec 2025

Not yet confirmed

• Solo Founder & AI Product Engineer Independently founded and bootstrapped a self-directed AI venture to research, build, and prototype production-grade autonomous decision-making systems and multi-agent workflows.
• Spearheaded end-to-end product strategy, full-stack software development, and scalable cloud infrastructure design.
• Successfully engineered and validated high-performance technical proof-of-concepts (POCs) before pausing execution due to funding constraints.
• 🛠️ Deployed Product Portfolios
• 1.
• Autonomous Due Diligence Agent
• System Design: Built a multi-agent investment research platform leveraging LangGraph and Hybrid RAG across 10,000+ SEC filings to automate complex financial, legal, and market-risk assessment workflows.
• State & Memory: Implemented stateful agent memory, workflow persistence layers, risk-scoring frameworks, and real-time analyst dashboards to track data lineage across long-running research loops.
• Cross-Domain Logic: Developed cross-domain reasoning workflows where financial analysis agents dynamically pass structured context to separate legal evaluation nodes.
• 2.
• Fresh Prints ERP (Proof-of-Concept)
• Core Automation: Developed an AI-native ERP suite prototype integrating GPT-4o Vision, multi-agent workflows, logistics optimization, and business process orchestration.
• Multimodal Pipelines: Programmed visual manufacturing anomaly detection and design validation workflows using multimodal vision models.
• HITL Integration: Created operational monitoring dashboards with integrated Human-in-the-Loop (HITL) approval gates for real-time shipment planning and procurement tracking.

S

Founder & Al Product Engineer

Startup Venture · Jul 2025 – Dec 2025

Not yet confirmed

• Founded an early-stage Al startup; designed and built production-grade autonomous decision-making systems and multi-agent architectures from ground up.

U

Preparation & Research

UPSC Civil Services Examination · Oct 2018 – May 2025

Not yet confirmed

• National-Level Competitive Distinction: Qualified the rigorous multi-tier UPSC Civil Services Examination (Preliminary and written Mains stages) to reach the prestigious Interview (Personality Test) stage, ranking in the top 0.1 percentile out of a million candidates nationwide-demonstrating exceptional intellectual horsepower, analytical reasoning, and high-stakes execution.
• Advanced Quantitative Aptitude & Analytical Problem Solving: Mastered advanced mathematical modeling, data interpretation, logical deduction, and structured reasoning under strict time constraints, consistently scoring in top percentiles on analytical and quantitative aptitude evaluations (CSAT).
• Digital Public Infrastructure, Al Ethics & Technology Policy: Evaluated national-scale digital ecosystems (India Stack, UPI, Open Digital Ecosystems), data privacy and governance frameworks (DPDP Act), and Al ethics, analyzing scalable technological interventions to address multi-dimensional public challenges.
• Knowledge Systems, Information Retrieval & Self-Directed Research: Engineered structured personal knowledge bases and multi-disciplinary indexing workflows; maintained a rigorous 60+ hour/week research regimen cultivating extraordinary discipline, cognitive endurance, and complex systems thinking.

Data Scientist / Machine Learning Engineer

Accenture · Jul 2016 – Sep 2018

Not yet confirmed

• Developed ML and NLP pipelines for enterprise document processing, text classification, and named-entity recognition (NER).
• Built predictive models and feature-engineering pipelines using Python, Scikit-Learn, and deep learning architectures.
• Engineered invoice-document matching and reconciliation algorithms utilizing fuzzy string matching and vector similarity metrics, improving processing throughput.
• Constructed analytical dashboards and model monitoring visualizations to evaluate inference performance and drive business decisions.

Data Scientist

Accenture · Jul 2016 – Sep 2018

Not yet confirmed

• Data Scientist / ML Engineer Developed and deployed enterprise-grade machine learning and natural language processing (NLP) solutions for high-volume document intelligence, automated processing, and data analytics pipelines.
• ⚡ Key Contributions & Technical Deliverables
• Document Intelligence: Designed and engineered core NLP and machine learning pipelines for unstructured document ingestion, text classification, and structured named entity recognition (NER).
• Predictive Modeling: Developed end-to-end predictive models and statistical feature-engineering pipelines using Python, Scikit-Learn, and deep learning techniques.
• Algorithmic Reconciliation: Built automated invoice-to-purchase-order matching and financial reconciliation engines using fuzzy text matching and string similarity algorithms, significantly reducing manual validation overhead.
• Analytics & Observability: Created interactive data dashboards and model telemetry visualizations to track system performance, data drift, and key operational efficiency metrics to support data-driven executive decisions.

What drives their work

The Inventor

Conceives and constructs novel AI solutions from the ground up.

Shasank's superpower

You excel at pioneering and building complex, robust AI systems, transforming nascent ideas into functional, scalable products.

How Shasank works

You are best suited for teams focused on developing and deploying cutting-edge AI solutions, particularly those involving agentic systems and large-scale data processing.