AG / PORTFOLIO
HOM-00SHEET 00 / 08
Open to opportunitiesMiami, FL

Anish
Guntreddi.

building intelligent systems that actually ship.

Software & ML Engineer

ex-Walmart Global Tech · Georgia Tech MSCS '28 · Co-Founder & CTO, VibeLink · Co-Founder, Synaptix

I build production ML systems and the backend, systems-level, and quantitative infrastructure they run on — from a sub-100 ms EMG inference pipeline to a social app with 800+ users.

Applied AI / MLSystems / BackendQuantitative ComputingProduct / Research
View Work

Four versions — master, software, ML/AI, quant. Master CV (PDF) ↗

BLD · SHEET 0101 / 08

Currently building.

Two live ventures, running at the same time, pulling from opposite ends of the same skill set.

BLD-001·CTO & CO-FOUNDER
REV.A

VibeLink

Find your people — the communities and events around you.

Most campus and city communities are invisible until you already know someone inside them. VibeLink surfaces the organizations, communities, and events around you and gets you in the door. As CTO and co-founder I own the product end to end — the React Native app, the FastAPI and PostgreSQL (Supabase) backend, authentication, notifications, and deployment — and took it from an idea to 800+ users and 70+ partner organizations, with a 4.8-star App Store rating.

try-vibelink.com
800+ users70+ organizations4.8★ App Store ratingCane Angel People's Choice Award

AREAS

mobile engineeringbackend systemsproduct architecturerecommendation systemsanalyticscloud infrastructureapplied AI
BLD-002·CO-FOUNDER & FOUNDING ML ENGINEER
REV.A

Synaptix

EMG biosignals, decoded into real-time control.

Rehabilitation devices mostly run fixed programs. Synaptix reads a patient's muscle activity (EMG) and adapts assistance in real time. I lead the ML work: the Python/PyTorch pipeline for acquisition, denoising, segmentation, feature construction, training, and evaluation — inference profiled under 100 ms, movement classification ~30% better than a rule-based baseline — integrated with FastAPI, Kotlin, and BLE/Arduino device control. It started as an independent project; it now advances through the University of Miami's USTAAR inventor program toward a patent filing, and won the Inventor's Ignite Award.

<100 ms inference~30% lift vs. rule-based baselineUniversity of Miami–backedAdvancing toward patent filingInventor's Ignite Award (USTAAR)

AREAS

machine learningEMG signal processingreal-time inferencecomputer visionembedded/device integrationHCIresearch engineering

“VibeLink is engineering for people at scale — product decisions, reliability, real users, constant iteration. Synaptix is engineering under uncertainty — noisy signals, models, and inference on a hard real-time budget. Same engineer, both ends of the spectrum: from open research to systems people actually use.”

FLG · SHEET 0303 / 08

Flagship work.

Five projects that carry the most weight — each one proving a distinct capability, from shipped products to systems built from first principles.

FLG-001·FLAGSHIP
REV.A

VibeLink

What it proves

Ships products people actually use

Social discovery for communities and events — 800+ users, 70+ partner organizations, a 4.8-star App Store rating, and the Cane Angel People's Choice Award. I own the mobile app, backend, auth, notifications, and deployment.

React NativeTypeScriptFastAPIPostgreSQL
View projectSTATUS · LIVE
FLG-002·FLAGSHIP
REV.A

Synaptix

What it proves

Real-time ML on noisy biosignals

An EMG pipeline that infers patient intent in under 100 ms and beats a rule-based baseline by ~30% — now a University of Miami–backed project advancing toward a patent filing, and winner of the Inventor's Ignite Award.

PyTorchSignal ProcessingEmbedded SystemsReal-time Inference
In active developmentSTATUS · BUILDING
FLG-003·FLAGSHIP
REV.A

FaultSmith

What it proves

AI engineering with security and evaluation rigor

Deliberate-debugging labs powered by LLM agents: validated failure injection, sandboxed execution, rate limiting, a written threat model, and deterministic fallbacks for when the model is wrong.

LLM AgentsSandboxingThreat ModelingEvaluation
View projectSTATUS · LIVE
FLG-004·FLAGSHIP
REV.A

Systems Engineering Portfolio

What it proves

C/C++ and full-stack fundamentals, 700+ tests

Five full-stack apps and five C/C++ systems projects — an 11.3M ops/sec matching engine, an HTTP server from raw TCP, a Redis-compatible cache, a Unix shell, and a memory allocator — each with automated tests.

C++17C11FastAPIReact/TSPostgreSQL
View projectSTATUS · LIVE
FLG-005·FLAGSHIP
REV.A

LLM Engineering Portfolio

What it proves

LLM specialization with experimental rigor

QLoRA fine-tuning, RAG evaluation, and sandboxed agent execution, with reproducible before/after metrics for every experiment.

QLoRARAGAgent SandboxingEvaluation Harnesses
View projectSTATUS · LIVE
PRF · SHEET

Proof, not just projects

External validation, not self-claims.

0+Product users
0+Partner organizations
0/800+Hackathon rankWalmart Jirafy hackathon, Top 10 of 800+ entries
0+Automated tests
0+Enterprise documents processed
<0msML inference latency
~0%Agent task-success improvement
EXP · SHEET 0101 / 08

Where I've shipped.

Industry experience and academic research across AI, quantitative finance, and full-stack systems. Two clusters below: industry assignments and research appointments.

EXP.A
N=6

Founder, consulting, internship, and applied-AI research engineering across enterprise AI, coding agents, quantitative analysis, and full-stack product.

IND-001·JANUARY 2026 – PRESENT
REV.A

VibeLink

CTO & Co-Founder

800+ users, 70+ organizations, 4.8-star App Store rating, Cane Angel People's Choice Award

  • Built and scaled a production social and events platform to 800+ users and 70+ organizations, owning features across mobile UI, backend APIs, data modeling, authentication, notifications, and deployment
  • Designed and shipped end-to-end product workflows across React Native, TypeScript, FastAPI, and PostgreSQL while balancing usability, reliability, maintainability, and rapid iteration
  • Turned feedback from users and organizations into shipped product improvements and production fixes; the product earned the Cane Angel People's Choice Award and a 4.8-star App Store rating
IND-002·APRIL 2026 – AUGUST 2026
REV.A

Veritas Prime

AI Consultant Engineer / Process Engineering Consultant

Enterprise agentic workflows shipped with validation, monitoring, and human review built in

  • Designed and shipped enterprise agentic software with Python and TypeScript, LangChain/LangGraph, RAG, APIs, retrieval, planning, and structured tool execution, owning architecture through deployment and iteration
  • Built reusable integrations and workflow components across SAP SuccessFactors and Salesforce, translating ambiguous stakeholder requirements into maintainable services, prototypes, and production automations
  • Added validation, logging, monitoring, retries, recovery, regression checks, and human review to surface failure modes and improve reliability, observability, and confidence in production AI workflows
IND-003·SEPTEMBER 2025 – MARCH 2026
REV.A

Mercor

Applied AI Researcher & Software Engineer, Coding Agents

20–25% higher solution acceptance and ~40% more automated test coverage for coding-agent evaluation

  • Built evaluation infrastructure for coding agents across production-style Python/Linux repositories using execution traces, behavioral tests, benchmark harnesses, CI automation, regression suites, and structured failure taxonomies
  • Reproduced software failures, debugged agent-generated code, and converted recurring failure modes into measurable tests and evaluation signals, improving solution acceptance by 20–25%
  • Expanded automated test coverage by ~40% through systematic debugging, profiling, code review, CI analysis, and regression testing while documenting findings for repeatable engineering decisions
IND-004·JUNE 2025 – AUGUST 2025
REV.A

Walmart Global Technology

AI Engineering Intern

~25% improvement in end-to-end task success across 5,000+ documents; Jirafy hackathon Top 10 of 800+ entries

  • Built enterprise AI applications with Python/FastAPI, Next.js, LangChain/LangGraph, Milvus, embeddings, and vector retrieval across 5,000+ structured and unstructured documents, improving end-to-end task success by ~25%
  • Deployed services with Docker, Kubernetes, CI/CD, logging, and monitoring, supporting scalable internal workflows and production-quality operation across data, retrieval, and application layers
  • Evaluated task success, retrieval quality, latency, consistency, and failure modes; partnered with engineering and product teams to incorporate feedback and iterate on reliable user-facing workflows
IND-005·JUNE 2024 – AUGUST 2024
REV.A

DUCAB

Quantitative Analyst & Data Science Intern

~60% less manual processing, ~20% better execution efficiency, ~50% fewer operational failures

  • Analyzed copper and aluminum derivatives, pricing/reference data, hedging workflows, and trade-monitoring processes tied to LME markets and execution systems
  • Built Python, Excel VBA, API, and SAP-connected automations that reduced manual processing by ~60% and improved execution efficiency by ~20%
  • Implemented monitoring, risk, and stop-loss workflow improvements that reduced operational failures by ~50% while translating market and business constraints into reliable quantitative tooling
IND-006·AUGUST 2023 – DECEMBER 2023
REV.A

AVAS Technology

Full-Stack Software Engineering Intern

Customer-facing web features shipped end to end, from REST endpoints to responsive UI

  • Built and maintained customer-facing web features using PHP, Laravel, JavaScript, MariaDB, and Bootstrap, spanning backend routes, REST endpoints, database models, and responsive user interfaces
  • Refactored backend and data workflows, wrote SQL queries, and debugged frontend, server, and database issues while validating releases against functional requirements
  • Added unit tests, participated in code review, documented changes, and troubleshot user-facing issues through iterative development and release cycles
EXP.B
N=3

Academic and applied research in clinical ML, software methodology, and EMG biosignal decoding through Synaptix, a University of Miami–backed project.

RES-001·APRIL 2025 – AUGUST 2025
REV.A

University of Miami Mathematical Finance Lab

Machine Learning Researcher

~60% directional accuracy vs. ~52% baseline across ~100 experiments; Beyond the Books Research Scholar

  • Applied Transformers, HMMs, GARCH/EGARCH, Monte Carlo methods, and statistical learning to OHLCV, quote/trade, volatility, and derivatives data across quantitative research pipelines
  • Ran ~100 walk-forward and out-of-sample experiments and evaluated Sharpe, Sortino, CAGR, maximum drawdown, hit rate, transaction costs, and regime sensitivity; strongest models reached ~60% directional accuracy versus a ~52% baseline
  • Co-authored mathematical-finance research supported by the Beyond the Books Research Scholarship and presented through the University of Miami Research Gallery
RES-002·MARCH 2025 – MAY 2026
REV.A

University of Miami Orthopedics Research Lab

Software Engineering Researcher

BLE-connected EMG research software with offline workflows and structured session logging

  • Developed Kotlin Android software for BLE-connected EMG/Arduino systems with real-time sensor ingestion, local persistence, offline workflows, synchronization, and structured session logging for biomedical research
  • Partnered with biomedical researchers to translate evolving study requirements into maintainable software and repeatable workflows while communicating technical tradeoffs across software and hardware boundaries
  • Debugged device communication, state recovery, synchronization, and hardware-software failures through iterative testing and root-cause analysis, improving reliability of connected research workflows
RES-003·OCTOBER 2025 – PRESENT
REV.A

Synaptix (University of Miami–backed)

Co-Founder & Founding ML Engineer

~30% better movement classification vs. rule-based baseline, sub-100 ms inference; Inventor's Ignite Award (USTAAR)

  • Co-founded Synaptix, an Inventor's Ignite Award-winning USTAAR neuromuscular-interface project advancing through the University of Miami inventor program toward patent filing; lead the ML work for physiological-signal classification and real-time inference
  • Built the Python/PyTorch pipeline for acquisition, denoising, segmentation, feature construction, training, and evaluation; improved movement classification by ~30% over a rule-based baseline and profiled sub-100ms inference
  • Integrated ML with FastAPI, Kotlin, BLE/Arduino, and device-control software while debugging data-quality, timing, communication, and cross-layer reliability failures across the end-to-end system
PRJ · SHEET 0202 / 08

Full range.

Twelve deep-dive portfolios backing up the four spikes.

Applied AI / ML

Quantitative Computing

Systems / Backend

Supporting Capabilities

iOS PortfolioAndroid PortfolioSecurity Engineering PortfolioCloud & DevOps PortfolioData Engineering PortfolioDeveloper Tools PortfolioAccessibility PortfolioCapstone Portfolio
SKL · SHEET 0303 / 08

Parts inventory.

A working inventory of languages, frameworks, and tooling — 108 parts across 5 bins, organized around four core spikes plus supporting capabilities. Hover a chip to inspect; everything here has been shipped to production or used in a research artifact.

SKL-A·N=18
REV.A

Applied AI / ML

BIN A
PyTorchTensorFlow / KerasHugging Facescikit-learnTransformersCNNsHMMsComputer VisionTime SeriesLangChainLangGraphRAGEmbeddings & Vector RetrievalLoRA / PEFTAgent EvaluationHuman-in-the-loop WorkflowsReal-time InferenceSignal Processing
SKL-B·N=28
REV.A

Systems / Backend

BIN B
PythonTypeScriptJavaScriptJavaC++17CC#KotlinSQLFastAPIFlaskREST APIsLinux / UnixTCP/IPConcurrency & MultithreadingDockerKubernetesAWSGCPPostgreSQLMongoDBRedisSupabaseFirebaseMilvusCI/CDAutomated TestingProfiling
SKL-C·N=17
REV.A

Quantitative Computing

BIN C
Probability & StatisticsLinear RegressionTime SeriesGARCH / EGARCHMonte CarloWalk-forward ValidationCalibrationMarket MicrostructureOrder BooksBacktestingPortfolio AccountingRisk MetricsRNumPypandasstatsmodelspybind11
SKL-D·N=10
REV.A

Product / Research

BIN D
ReactNext.jsReact NativeSystem DesignDesign PatternsCode ReviewGitRegression TestingClaude CodeCodex
Supporting CapabilitiesN=35
SwiftSwiftUICombineCore DataJetpack ComposeCoroutinesRoomHiltXCTestMobile DevelopmentTerraformAzureGitHub ActionsPrometheusVercelNetlifyCloud & DevOpsPenetration TestingThreat ModelingVulnerability ResearchBurp SuiteMetasploitWiresharkOWASPSecurity EngineeringApache SparkAirflowdbtSnowflakeKafkaData EngineeringWCAG 2.2ARIAaxe-coreWeb Accessibility
EDU · SHEET 0404 / 08

Academic record.

Graduate study in computer science at Georgia Tech (thesis option, 4.0 GPA), built on a triple-major foundation in computer science, mathematics, and AI at the University of Miami as a Stamps Scholar.

EDU-001·GEORGIA INSTITUTE OF TECHNOLOGY
REV.A

Georgia Institute of Technology

Atlanta, GA · EXPECTED MAY 2028

  • 01.M.S. Computer Science

Thesis option · GPA 4.0

EDU-002·UNIVERSITY OF MIAMI
REV.A

University of Miami

Miami, FL · MAY 2026

  • 01.B.S. Computer Science
  • 02.B.S. Data Science and Artificial Intelligence
  • 03.B.S. Mathematics

Triple major · Stamps Scholar · Foote Fellows Honors Program

RELEVANT COURSEWORKN=10
Data Structures & AlgorithmsMachine LearningDeep LearningDatabase SystemsSoftware EngineeringOperating SystemsComputer NetworksFinancial MarketsQuantitative FinanceStatistical Methods
  1. 01.

    Stamps Scholar

    University of Miami

    Selected for the highly selective Stamps Scholarship based on academic achievement, leadership, service, and potential; part of the national Stamps Scholars community.

  2. 02.

    Presidential Scholarship

    University of Miami

    Merit scholarship recognizing strong academic achievement and continued scholastic performance.

  3. 03.

    Foote Fellows Honors Program

    University of Miami

    Honors program for high-achieving undergraduates pursuing an academically ambitious, interdisciplinary course of study.

  4. 04.

    Inventor's Ignite Award

    USTAAR, University of Miami

    Awarded to Synaptix through the University of Miami USTAAR inventor program, supporting continued development of the neuromuscular-interface invention and the university patent/IP process.

  5. 05.

    Beyond the Books Research Scholar

    University of Miami

    Competitive faculty-mentored research scholarship supporting mathematical-finance research added to the University of Miami Research Gallery.

  6. 06.

    Walmart Jirafy Hackathon — Top 10 of 800+

    Walmart Global Technology

    Jirafy, an AI-assisted JIRA workflow prototype, improved engineering productivity and reduced ticket-handling time in project evaluation.

  7. 07.

    Cane Angel People's Choice Award

    University of Miami entrepreneurship ecosystem

    Recognized VibeLink for strong community support for the product.

HBY · SHEET 0505 / 08

Beyond the code.

What I do when I'm not shipping code or training models — the unofficial, unoptimized part of the stack.

WNO.01

Weightlifting

I got into weightlifting when I was 18 to balance out my life with a good healthy habit on top of going to classes and working on projects.

SINCE AGE 18
GNO.02

Gaming

I have been gaming since I was 7 years old and continue to try new games every day. My favorite genres are action, horror, sports, and indie.

SINCE AGE 7
TNO.03

Tennis

I have been playing tennis since I was 9 years old and was a varsity athlete in high school.

SINCE AGE 9 · VARSITY
FNO.04

Football

Love to play football for fun with my friends.

FOR FUN
BNO.05

Basketball

Started playing a lot more recently with my friends.

RECENTLY PICKED UP
CNO.06

Cooking

Got into cooking the same time as weightlifting to begin meal prepping after workouts.

POST-WORKOUT MEALS
HNO.07

Hanging with Friends

Love to try new activities with my friends.

ALWAYS TRYING NEW
TRM · SHEET 0606 / 08

Interactive shell.

A live REPL with the same data this page is built from. Type `help` for commands, `tree` for structure, `whoami` for the fast version.

anish@portfolio:~PTY · 80×24
anish@dev ~ terminal
Welcome! Type 'help' to see available commands.
anish@dev ❯
CTC · SHEET 0707 / 08

Open channels.

Email is fastest. GitHub for code, LinkedIn for everything else — or use the form below; it goes straight to my inbox.

CTC-001·EMAIL
REV.A
↗ OPEN
EMAIL
anishguntreddi@gmail.com
CTC-002·GITHUB
REV.A
↗ OPEN
GITHUB
github.com/Anish-Guntreddi
CTC-003·LINKEDIN
REV.A
↗ OPEN
LINKEDIN
linkedin.com/in/anish-guntreddi
CTC-FORM·SEND A MESSAGE
REV.A
NAMEREQUIRED
COMPANY
TYPE
SUBJECTREQUIRED
MESSAGEREQUIRED
GOES STRAIGHT TO MY INBOX