RAMCHAND KUMARESAN
PMP, PgMP
AI Researcher & Program Leader | LLM Systems · On-Device Inference | Enterprise-Scale Delivery
Professional Summary
AI researcher and program leader who pairs hands-on machine-learning research with a 16-year record of enterprise-scale delivery. Author of six research papers (arXiv, NeurIPS 2026, TMLR) on large-language-model systems — spanning direct Apple Neural Engine programming, cooperative model training, multi-model orchestration, evolutionary fine-tuning, and pre-registered methodology for screening evolutionary methods — with headline results including a +21.76% cross-lingual model-fusion gain and an 8.5× Apple Neural Engine compilation speedup, every result reproducible from committed code. Builder of production AI systems, from a live personalized-storytelling platform to a clinical-triage assistant that runs a 5GB language model fully offline on a phone. This research depth sits on a foundation of dual PMI certifications (PMP, PgMP) and $25M+ multi-state technology programs delivered on time across cloud, SaaS, and government domains. Equally fluent translating ambiguity into measurable business outcomes — KPIs, OKRs, ROI, audit-ready delivery — and translating a research hypothesis into a reproducible, pre-registered experiment.
Experience
Founder & AI Researcher
Murai Labs
2025 — Present
Austin, TX
- Authored six AI research papers (arXiv, NeurIPS 2026, TMLR) on large-language-model systems — spanning direct Apple Neural Engine programming, cooperative model training, multi-model orchestration, evolutionary mixture-of-LoRA fine-tuning, and pre-registered cheap-baseline screening for evolutionary methods.
- Shipped production AI products end-to-end: StoryForge Kids, a live iOS + web personalized-storytelling platform, and Marunthagam, a Tamil-first clinical-triage assistant running a 5GB language model fully offline on an Android phone (0/12 missed emergencies, 100% safety-refusal on adversarial prompts).
- Delivered headline research results — +21.76% cross-lingual model-fusion gain (KALAVAI) and 8.5× Apple Neural Engine compilation speedup (Orion) — each reproducible from committed code and audit-tracked artifacts.
- Built deep hands-on command of the modern AI stack: PyTorch, HuggingFace, LoRA/QLoRA fine-tuning, mixture-of-experts routing, quantization, and edge deployment across CUDA, Apple Metal, and the Neural Engine.
- Established disciplined research methodology — pre-registered hypotheses, frozen-hash provenance, and cheap-baseline falsification — publishing negative results with the same rigor as positive ones.
Enterprise Project Manager
Procore Technologies
Dec 2025–Present
Austin, TX (Remote)
- Manage enterprise project delivery for public sector and government clients.
- Lead cross-functional coordination across engineering, product, customer success, and compliance.
- Drive change management using SDLC methodology — zero disruption to government operations.
- Partner with Change Advisory Board (CAB) for FedRAMP/SOC 2 compliance alignment.
- Facilitate Agile sprint planning for distributed teams — 15% reduction in release cycle time.
Senior Project Manager, Program Management Office
Aspira
2017 — 2025
Dallas, TX
- Led multi-year, multi-state modernization programs (Utah, Texas, Kansas, NJ, Alberta, Maryland) valued over $25M, delivering SaaS platforms for park reservations, licensing, and customer engagement.
- Directed enterprise API integrations connecting DMV, DCF, banking, and payment systems (NIC Pay) to core platforms, enabling real-time data exchange and supporting $40M+ revenue in year one.
- Led Aspira's enterprise migration from physical datacenters to AWS (EC2, S3, Redshift), partnering with DevOps, data, and security teams to modernize infrastructure, reduce costs by 30%, and achieve 99.99% uptime.
- Managed large-scale data migrations for 600K+ customer records across 28 parks (5,000+ sites) with zero data loss and seamless cutover to cloud environments.
- Spearheaded deployment of AI-driven conversational chatbots, automating Tier-1 customer interactions and achieving 35% reduction in call volume and $250K+ annual support cost savings.
- Served as primary liaison for state agencies, business sponsors, and technical teams—reporting project progress, managing risks, and maintaining alignment across multiple integration streams.
- Mentored project managers and established Project Management Office by standardizing project and program management approaches.
Senior Project Manager
Doodleblue Innovations Ltd
2014 — 2017
Chennai, India
- Delivered mobile-first digital solutions in iOS and Android environments, including customer self-service and support apps for clients across finance, retail, and services.
- Established the company's presence in the HR industry and founded an in-house product development division.
- Mentored and trained 10+ project managers, instilling best practices in initiation, change management, and stakeholder engagement.
- Increased organizational efficiency through the creation and execution of project and portfolio management procedures.
- Acted as delivery lead for app development projects, ensuring on-time delivery and quality assurance across distributed teams.
Enterprise Release Project Manager
Cognizant Technology Solutions
2012 — 2014
Hartford, CT
- Supervised enterprise application releases and updates using SDLC methodology, ensuring seamless deployment with minimal impact on critical domains like Real-Time Eligibility and Claims Adjudication.
- Drove continuous delivery for pharmacy claims adjudication systems and improved team efficiency using Agile methodology.
- Coordinated with cross-functional teams including QA, UAT, development, and project managers for timely production releases.
- Optimized release management processes by directing a team of 3 release managers.
RF Project Lead
Glow Networks Inc
2011 — 2012
Dallas, TX
Case Lab Business Analyst
Florida A&M University
2009 — 2010
Tallahassee, FL
AI Research & Publications
Orion
arXiv 2603.06728 ↗Programming Apple's Neural Engine — the first open end-to-end system for direct ANE inference and training.
↳ 8.5× ANE compilation speedup; bypasses CoreML entirely.
Predicting when independent specialist fusion works in cooperative LLM training.
↳ +21.76% cross-lingual fusion gain; fusion gain predicted before compute is spent.
Adaptive complexity routing for multi-model ensembles via self-consistency variance.
↳ Evaluated on 1,510 tasks across 4 benchmarks with 7,550+ auditable runs.
Decomposing evolutionary mixture-of-LoRA architectures with a controlled factorial.
↳ Rigorous attribution + honest-negative-result methodology (pre-registered, audit-chained).
Rigel
arXiv 2606.12765 ↗Reverse-engineering the Apple M4 Max Metal 4.1 tensor compute path.
↳ 11 hidden hardware specifications recovered; +12.9% hand-fused GEMM kernel.
MARMAM
arXiv 2606.29119 ↗A pre-registered cheap-baseline screening rule for knowing when an evolutionary outer loop will not help.
↳ A ~50–70 GPU-hour gate that prevented an estimated 400+ GPU-hours of needless compute.
Key Projects
StoryForge Kids
Live personalized children's storytelling platform (iOS + web) with a multi-pass story engine and persistent character memory.
↳ Shipped end-to-end across iOS and web — a production AI product with real users, not a prototype.
Marunthagam
Tamil-first clinical-triage assistant running a 5GB language model fully offline on an Android phone for rural community-health workers.
↳ 0/12 missed emergencies; 100% safety-refusal on adversarial prompts. Submitted to the Gemma hackathon.
PINNAL
RNA-therapeutic candidate ranker that prioritizes antisense-oligo interventions for splice-altering variants using an ensemble of genomic foundation models (Evo 2, Borzoi, Enformer, ViennaRNA).
↳ Pre-alpha, demo-ready. Surfaced all 3 tiles of the clinically-validated SMN2 ISS-N1 site in its top 10 (enrichment@10 4.10 vs 0.00 for Evo 2 used as a ranker) in a pre-registered falsification test.
KALAVAI
Cooperative LLM-training protocol that fuses independently trained specialists and predicts the fusion gain before any compute is spent.
↳ +21.76% cross-lingual gain; submitted to NeurIPS 2026 with a full reproducibility audit trail.
Multi-State SaaS Platform
Enterprise park reservation and licensing system deployed across Utah, Texas, Kansas, NJ, Alberta, and Maryland.
↳ $25M program delivered on time; zero disruption to state government operations across 6 jurisdictions.
Enterprise Cloud Migration
AWS migration (EC2, S3, Redshift) reducing costs 30% with 99.99% uptime.
↳ 30% infrastructure cost reduction; 99.99% uptime maintained post-cutover.
AI Conversational Chatbot
Tier-1 customer-interaction automation deployed at enterprise scale.
↳ 35% reduction in call volume; $250K+ in annual support cost savings.
Skills & Expertise
Certifications
Project Management Professional (PMP)
PMI
Program Management Professional (PgMP)
PMI
Business English Certification
Cambridge University — Class A
Education
Master of Science in Industrial Engineering & Management
Florida Agricultural and Mechanical University
2010
Bachelor of Engineering in Mechanical Engineering
Anna University, Chennai
2006