bg ▸ agent_workflow.sim()
// portfolio.init() Tampa · FL · 2026
// senior_software_engineer

SHIVAM SHARMA

Senior Software Engineer — building production Gen-AI & multi-agent systems.
Tampa, FL (open to relocation) · [email protected] · linkedin.com/in/shivamsharma00
01
Summary
Senior software engineer with 10+ years building production systems, now focused on Gen-AI and multi-agent applications. I design LLM-native systems where agents orchestrate at scale, outputs stay verifiable and auditable, and infrastructure degrades gracefully under failure. Deep enterprise experience across life sciences (Pfizer) and financial/enterprise (PwC), with end-to-end ownership from data pipeline through agent orchestration, LLM integration, and Kubernetes deployment.
02
Skills
Gen AI / LLM
Multi-agent orchestration, RAG, structured output, evals, MCP, prompt versioning, human-in-the-loop, LangChain, LangGraph, Langfuse, GPT-4o, embeddings
Languages
Python, TypeScript / JavaScript, SQL
Backend
FastAPI, FastStream, asyncio, Pydantic, REST, httpx
Data & Messaging
PostgreSQL, pgvector, MongoDB, Redis, RabbitMQ, Kafka, DynamoDB, S3
Cloud & DevOps
AWS (Lambda, API Gateway, Step Functions, ECS, S3, DynamoDB, CloudWatch), Azure (Entra ID, Key Vault, Event Hub), Docker, Kubernetes, Argo CD, Helm, GitHub Actions
ML / CV / NLP
PyTorch, OpenCV, MediaPipe, SAM2, SpaCy, NLTK, Tesseract
Frontend
Vue 3 (Composition API), Vite, SSE
03
Experience
Forward Deployed Engineer (Gen AI)
2025 – Present
PwC · Tampa, FL
Multi-agent threat-intelligence platform — Designed and shipped a production system that autonomously maps an organization's external attack surface (subsidiaries, domains, subdomains, IP ranges/ASNs, open ports), replacing manual reconnaissance with an LLM-augmented agent swarm.
Architected it as containerized microservices — a FastAPI supervisor orchestrating FastStream worker agents over RabbitMQ RPC, with PostgreSQL + pgvector and Redis — and designed a resilient dispatch layer that runs the identical pipeline distributed or in-process as a fallback, so a broker outage degrades gracefully instead of hard-failing the scan.
Built a deterministic-tools-first, LLM-as-gap-filler pattern with strict JSON-schema output, a JSON-repair layer, and Langfuse tracing/evals; streamed live progress to a Vue 3 SPA via SSE over Redis, and shipped to Kubernetes via Argo CD GitOps with a GitHub Actions pipeline building 7 services in parallel across four environments.
AI-assisted risk & compliance service — Built a FastAPI + MongoDB microservice that grounds risk and control generation in NIST 800-53 (~1,196 controls) using a deterministic-retrieval → LLM-judgment → human-in-the-loop pipeline, keeping every AI output auditor-defensible.
Developed a multimodal evidence-review pipeline (vision models for screenshots, text extraction for documents) that rates control effectiveness on a 0–100 scale, backed by an offline eval harness scoring AI output against ground truth; exposed the REST surface as MCP tools so the same functions power both the product and a tool-calling agent.
LLM access-review system — Built a LangChain/LangGraph workflow that ingests user-access signals, classifies security risk with AI-generated reasoning and deterministic fallback rules, and surfaces results to managers — owning the full stack from data pipeline through LLM integration and dashboard.
Senior Software Engineer (Tech Lead)
2024 – 2025
Pfizer
Health-literacy validation tool (CLEAR) — Designed and led a linguistic-validation system that automated proofreading of scientific content, applying 50+ NLP rule checks (SpaCy, NLTK, custom Python) across grammar, terminology, tone, and regulatory compliance.
Architected it as serverless microservices on AWS Lambda orchestrated with Step Functions — REST ingestion, async processing, S3 for outputs and DynamoDB for metadata and version tracking, with CloudWatch monitoring and alerting.
Document-to-deck generation tool — Built a system that converts white papers and technical documents into PowerPoint decks using GPT-4o for summarization and sectioning and python-pptx for dynamic slide generation, with image and table extraction via PyMuPDF, Tesseract, and OpenCV.
Data Scientist
Apr 2024
Flexon Technologies
Built ML and data-science pipelines in Python — data ingestion, feature engineering, and model development — delivering analyses and models into production workflows.
Machine Learning Engineer
Aug 2023 – Mar 2024
PinHous
Used computer vision to extract property features from real-estate images and built recommendation engines (collaborative + content-based filtering) that increased user interaction with recommended listings by ~20%.
Set up and optimized MySQL on AWS RDS and engineered data pipelines for personalized house feeds, and directed A/B testing aimed at lifting conversion rates.
Computer Vision Engineer
Sep 2019 – Aug 2021
ActiveIQ
Built deep-learning kits enabling autonomous functionality for commercial vehicles — developed a lane algorithm (gradient filters, perspective transform, custom segmentation in PyTorch) that reduced offset by 30%.
Achieved 96% pixel accuracy for assisted-driving mode on Indian roads and collaborated on sensor-fusion (camera, LiDAR, radar) strategies improving perception accuracy by ~20%.
04
Projects
Only — Personal AI agent
github →
A software-based personal AI built on a multi-agent harness: workflow orchestration, prompt versioning, MCP integrations, local models via Ollama, and a voice pipeline routing iPhone → MacBook over Tailscale. Python.
Only Notes — Cross-platform notes app
github →
Tauri + Vue 3 + SQLite desktop app with a GitHub Actions CI/CD pipeline and macOS notarization.
// research_&_academic
Bidirectional Search heuristics
github →
Bi-directional search with MM0 / MM heuristics, benchmarked for effectiveness across complex search spaces.
Authorship Obfuscation nlp
github →
Ensemble of SVM classifiers with the MUTANT-X obfuscator — high METEOR scores while cutting obfuscation time to ~1 minute.
Proximal Training for GANs generative
github →
Compared proximal training against other GAN methods, producing high-quality CelebA samples with Wasserstein GANs.
Deep Q-Learning — Autonomous Navigation rl
github →
Deep Q-learning navigation policy, simulating successful car movement in the CARLA simulator.
Camera Precision for Mono-SLAM perception
github →
Investigated how camera resolution and field of view affect MonoSLAM precision, with recommendations for optimal selection.
Collective Movement in Robot Swarms multi-robot
github →
Simulated collective movement of multiple robots in MATLAB using collective localization.
6-DOF Robotic Arm embedded
github →
Designed, 3D-printed and controlled a 6-DOF robotic arm with MATLAB + Arduino, improving kinematic accuracy.
05
Writing
medium.com/@shivamsharma00
06
Education
M.S., Robotics & Autonomous Systems (AI Specialization)
Arizona State University
2021 – 2023
B.Tech, Computer Science & Engineering
Rajasthan Technical University
2015 – 2019
07
Contact
// end_of_file — built in Tampa · galeetch.com