AI contractor for NASA centers and mission directorates.

Why NASA is buying AI in 2026

NASA's AI appetite has expanded across every mission directorate. On-board autonomy is mandatory for deep-space missions where light-speed delay kills teleoperation. Earth-observing constellations now produce more imagery in a day than analysts can review in a month, and the Agency is explicit that ML is the leverage. Mission operations, clinical support for long-duration crewed missions, spacecraft health monitoring, and science-archive analytics are all AI line items in the current budget cycle.

Precision Delivery Federal LLC is an Ames, Iowa-based small business aimed at NASA AI/ML scope. (Naming coincidence: our city is Ames, Iowa; NASA's Ames Research Center is in Moffett Field, California. Separate Ameses, both AI-relevant.) We are SAM.gov active, NAICS 541512 primary, SBIR-registered, and we carry a production federal ML past performance anchor at SAMHSA. Our stack spans frontier LLMs on FedRAMP paths, open-weight models for constrained environments, and scientific ML for the PDE-governed systems that dominate NASA work.

NASA centers we target

• Jet Propulsion Laboratory (JPL) — Robotic planetary exploration, deep-space missions, on-board autonomy, Earth-science instruments. Caltech-operated FFRDC with extensive small-business partnering.
• Ames Research Center (ARC) — AI, autonomy, air-traffic management, astrobiology. Heart of NASA's applied AI work.
• Goddard Space Flight Center (GSFC) — Earth observation, heliophysics, astrophysics, Landsat and MODIS data pipelines.
• Langley Research Center (LaRC) — Aeronautics, atmospheric sciences, entry-descent-landing systems.
• Glenn Research Center (GRC) — Propulsion, power systems, communications.
• Johnson Space Center (JSC) — Human spaceflight, crewed mission operations, life sciences and medical.
• Marshall Space Flight Center (MSFC) — Launch systems, propulsion, SLS and Artemis.
• Kennedy Space Center (KSC) — Launch operations, ground systems, payload processing.

NASA vehicles and opportunity paths

• NASA SBIR / STTR — Annual solicitations with AI/ML topics across every mission directorate (SMD, ARMD, STMD, SOMD, ESDMD). Post-April-2026 reauthorization the program is funded through 2031. We are DSIP-adjacent, SBIR-registered, and ready to submit.
• ROSES (Research Opportunities in Space and Earth Sciences) — SMD's omnibus research solicitation. AI/ML elements in most recent cycles, especially for Earth-observing data analytics.
• Center-specific BAAs and SAAs — JPL, Goddard, and Ames each run targeted calls. Small-business-friendly.
• SEAS (Solutions for Enterprise-Wide Procurement) — NASA's IT services IDIQ; AI/ML scope rides as task orders under prime holders.
• STMD BAA — Space Technology Mission Directorate's BAA funds technology maturation — including AI autonomy.
• SEWP VI via partners — hardware-adjacent AI deployments across centers.
• JPL subcontracts — JPL runs its own procurement system and contracts extensively with AI/ML small businesses.

AI/ML scope that fits NASA the best

• On-board autonomy — Decision-making for spacecraft, rovers, and landers where light-speed delay prevents human-in-the-loop control. Includes autonomous target selection, hazard avoidance, opportunistic science, and fault recovery.
• Earth observation ML — Land cover classification, deforestation and wildfire detection, flood and disaster response, methane and CO2 plume detection, agriculture monitoring. Our machine learning team is Kaggle-Top-200 caliber and built on computer-vision workloads of exactly this shape.
• Mission planning optimization — Constraint-satisfaction and reinforcement-learning approaches to observation scheduling, trajectory planning, and resource allocation under operational constraints.
• Spacecraft and instrument health — Anomaly detection on telemetry, predictive maintenance for ground systems and flight hardware.
• Astronaut decision support — Clinical ML for long-duration crewed missions, onboard knowledge retrieval, procedural assistance. This intersects directly with our SAMHSA clinical-data experience.
• Satellite constellation analytics — Cross-constellation fusion, gap-filling, super-resolution, and data-product assembly pipelines.
• Scientific ML over NASA archives — Foundation-model adaptation to astrophysics, heliophysics, and planetary datasets; retrieval-augmented research assistants for NASA scientists.
• Agentic AI for mission operations — Multi-agent systems that monitor telemetry streams, correlate with operational playbooks, and draft response actions for flight controllers with provenance and human-in-the-loop gating on every action.

Why a small business on NASA AI

NASA has real socio-economic goals and a culture that genuinely values small-business innovation — NASA SBIR is one of the most mature SBIR programs in the federal ecosystem, with Phase III commercialization pathways that actually close. A prime or FFRDC partnering with a SAM-registered, NAICS 541512 small business with AI/ML depth and production federal past performance gets specialized capability and small-business credit in one move.

NASA technologists also move fast. A JPL principal investigator working against an SBIR Phase II technical volume or a Goddard science team drafting a ROSES response does not have capacity to wait on a big prime's capture committee. We respond in days, not weeks.

Past performance and what we bring

Our confirmed federal past performance is a production ML system at SAMHSA (HHS) — a live system inside a federal ATO boundary, with NIST 800-53 controls, federal security review, and ongoing operations. That discipline is directly transferable to NASA ATO environments. Pre-Precision, Bo delivered cloud migration and data platform engineering at federal consulting firms supporting multiple civilian agencies, and competes as a Kaggle Top 200 data scientist — the computer-vision and ML competition muscle that NASA Earth-observing and planetary-imagery work demands.

For NASA-specific scope, we target and pursue rather than claiming delivered past performance at any NASA center. That distinction matters, and we will always state it clearly.

Stack for NASA workloads

• Computer vision — Modern transformer-based backbones (ViT, Swin, DINO-v2), segmentation (SAM-class), detection (DETR, YOLO-family), super-resolution, self-supervised pretraining on unlabeled remote-sensing corpora.
• Scientific ML — Physics-informed neural networks, neural operators (FNO, DeepONet), graph neural networks for orbital-mechanics and molecular work.
• Autonomy — Reinforcement learning, model-predictive control augmented by learned dynamics, constraint-satisfaction solvers.
• Foundation models — Claude, GPT-4, Llama, Mistral on appropriate federal paths; parameter-efficient fine-tuning on NASA-domain corpora.
• Data platforms — Lakehouse architectures for Earth-observation pipelines, experiment tracking, provenance tagging.
• Cloud and on-prem — AWS GovCloud, Azure Government, on-prem center HPC, hybrid burst patterns.
• Security — NIST 800-53 by default, audit logging, provenance on every generation, adversarial robustness testing.

Why Iowa for NASA

NASA Ames Research Center is in Moffett Field, California; we are Precision Delivery Federal in Ames, Iowa — a coincidence that makes for a memorable introduction. What Iowa actually gives NASA is geographic diversity in the small-business base, a strong Iowa State University computational ecosystem, and a vendor outside the coastal clusters that dominate most NASA AI contracting. For program managers who value diversification of their supplier base, we are a useful data point.

How to engage

If you are a NASA technical lead, a JPL PI, a center contracting officer, or a prime scoping an AI/ML subcontractor, email [email protected]. Include the center, the vehicle or topic number, and the scope. We respond within 24 hours with a fit assessment, rough level of effort, and a teaming construct.